(Modern Toxo Approaches

This PDF is available from The National Academies Press at http://www.nap.edu/catalog.php?
record_id=21775
Application of Modern Toxicology Approaches for Predicting

Acute Toxicity for Chemical Defense
ISBN
978-0-309-37666-2
130 pages
8.5 x 11
PAPERBACK (2015)
Committee on Predictive-Toxicology Approaches for Military Assessments

of Acute Exposures; Board on Environmental Studies and Toxicology;
Board on Life Sciences; Division on Earth and Life Studies; The National
Academies of Sciences, Engineering, and Medicine
Visit the National Academies Press online and register for...

Instant access to free PDF downloads of titles from the
NATIONAL ACADEMY OF SCIENCES
NATIONAL ACADEMY OF ENGINEERING
INSTITUTE OF MEDICINE
NATIONAL RESEARCH COUNCIL
10% off print titles

Custom notification of new releases in your field of interest
Special offers and discounts
Distribution, posting, or copying of this PDF is strictly prohibited without written permission of the National Academies Press.
Unless otherwise indicated, all materials in this PDF are copyrighted by the National Academy of Sciences.
Request reprint permission for this book
Copyright National Academy of Sciences. All rights reserved.
Application of Modern Toxicology Approaches for Predicting Acute Toxicity for Chemical Defense
PREPUBLICATION COPY
APPLICATION OF MODERN TOXICOLOGY

APPROACHES FOR PREDICTING ACUTE
TOXICITY FOR CHEMICAL DEFENSE
Committee on Predictive-Toxicology Approaches for

Military Assessments of Acute Exposures
Board on Environmental Studies and Toxicology
Board on Life Sciences
Division on Earth and Life Studies
THE NATIONAL ACADEMIES PRESS
500 Fifth Street, NW
Washington, DC 20001
This project was supported by Contract W81K04-11-D-0017, TO#8 between the National Academies of Sciences,
Engineering, and Medicine and the US Department of Defense. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the view of the organizations or agencies that provided support for this project.
Additional copies of this report are available for sale from the National Academies Press, 500 Fifth Street, NW,
Keck 360, Washington, DC 20001; (800) 624-6242 or (202) 334-3313; http://www.nap.edu.
Copyright 2015 by the National Academy of Sciences. All rights reserved.
Printed in the United States of America
Suggested citation: The National Academies of Sciences, Engineering, and Medicine. 2015. Application of Modern
Toxicology Approaches for Predicting Acute Toxicity for Chemical Defense. Washington, DC: The National Academies Press.
The National Academy of Sciences was established in 1863 by an Act of Congress, signed by President Lincoln, as a private, nongovernmental institution to advise the nation on issues related to science and technology.
Members are elected by their peers for outstanding contributions to research. Dr. Ralph J. Cicerone is president.
The National Academy of Engineering was established in 1964 under the charter of the National Academy of
Sciences to bring the practices of engineering to advising the nation. Members are elected by their peers for
extraordinary contributions to engineering. Dr. C. D. Mote, Jr., is president.
The National Academy of Medicine (formerly the Institute of Medicine) was established in 1970 under the
charter of the National Academy of Sciences to advise the nation on medical and health issues. Members are
elected by their peers for distinguished contributions to medicine and health. Dr. Victor J. Dzau is president.
The three Academies work together as the National Academies of Sciences, Engineering, and Medicine to provide independent, objective analysis and advice to the nation and conduct other activities to solve complex
problems and inform public policy decisions. The Academies also encourage education and research, recognize outstanding contributions to knowledge, and increase public understanding in matters of science, engineering, and medicine.
Learn more about the National Academies of Sciences, Engineering, and Medicine at www.nationalacademies.org.
COMMITTEE ON PREDICTIVE-TOXICOLOGY APPROACHES FOR

MILITARY ASSESSMENTS OF ACUTE EXPOSURES
Members
DAVID C. DORMAN (Chair), North Carolina State University, Raleigh, NC
WEIHSUEH A. CHIU, Texas A&M University, College Station, TX
HAIYAN HUANG, University of California, Berkeley, CA
ANDY NONG, Health Canada, Ottawa, ON
GRACE PATLEWICZ, US Environmental Protection Agency, Research Triangle Park, NC
DAVID REIF, North Carolina State University, Raleigh, NC
JOHN WADE, Battelle, Arlington, VA
KATRINA WATERS, Pacific Northwest National Laboratories, Richland, WA
BARBARA A. WETMORE, The Hamner Institutes for Health Sciences, Research Triangle Park, NC
YVONNE WILL, Pfizer, Groton, CT
Staff
ELLEN K. MANTUS, Project Director
MARILEE SHELTON-DAVENPORT, Senior Program Officer
KERI STOEVER, Research Associate
NORMAN GROSSBLATT, Senior Editor
MIRSADA KARALIC-LONCAREVIC, Manager, Technical Information Center
RADIAH ROSE-CRAWFORD, Manager, Editorial Projects
IVORY CLARKE, Senior Program Assistant
Sponsor
US DEPARTMENT OF DEFENSE
Prepublication copy
BOARD ON ENVIRONMENTAL STUDIES AND TOXICOLOGY1

Members
ROGENE F. HENDERSON (Chair), Lovelace Respiratory Research Institute, Albuquerque, NM
PRAVEEN AMAR, Independent Consultant, Lexington, MA
RICHARD A. BECKER, American Chemistry Council, Washington, DC
MICHAEL J. BRADLEY, M.J. Bradley & Associates, Concord, MA
JONATHAN Z. CANNON, University of Virginia, Charlottesville, VA
GAIL CHARNLEY ELLIOTT, HealthRisk Strategies, Washington, DC
DOMINIC M. DITORO, University of Delaware, Newark, DE
DAVID C. DORMAN, North Carolina State University, Raleigh, NC
CHARLES T. DRISCOLL, JR., Syracuse University, Syracuse, NY
WILLIAM H. FARLAND, Colorado State University, Fort Collins, CO
LYNN R. GOLDMAN, The George Washington University, Washington, DC
LINDA E. GREER, Natural Resources Defense Council, Washington, DC
WILLIAM E. HALPERIN, University of Medicine and Dentistry of New Jersey, Newark, NJ
STEVEN P. HAMBURG, Environmental Defense Fund, New York, NY
ROBERT A. HIATT, University of California, San Francisco, CA
PHILIP K. HOPKE, Clarkson University, Potsdam, NY
SAMEUL KACEW, University of Ottawa, Ontario
H. SCOTT MATTHEWS, Carnegie Mellon University, Pittsburgh, PA
THOMAS E. MCKONE, University of California, Berkeley, CA
TERRY L. MEDLEY, E. I. du Pont de Nemours & Company, Wilmington, DE
JANA MILFORD, University of Colorado at Boulder, Boulder, CA
MARK A. RATNER, Northwestern University, Evanston, IL
JOAN B. ROSE, Michigan State University, East Lansing, MI
GINA M. SOLOMON, California Environmental Protection Agency, Sacramento, CA
PETER S. THORNE, University of Iowa, Iowa City, IA
JOYCE S. TSUJI, Exponent, Inc., Bellevue, WA
Senior Staff
JAMES J. REISA, Senior Director
ELLEN K. MANTUS, Scholar and Director of Risk Assessment
RAYMOND A. WASSEL, Scholar and Director of Environmental Studies
DAVID J. POLICANSKY, Scholar
SUSAN N.J. MARTEL, Senior Program Officer for Toxicology
MIRSADA KARALIC-LONCAREVIC, Manager, Technical Information Center
RADIAH ROSE-CRAWFORD, Manager, Editorial Projects
This study was planned, overseen, and supported by the Board on Environmental Studies and Toxicology.
vi
Prepublication copy
OTHER REPORTS OF THE

BOARD ON ENVIRONMENTAL STUDIES AND TOXICOLOGY
Review of Californias Risk-Assessment Process for Pesticides (2015)
Sustainability Concepts in Decision Making, Tools and Approaches for the US Environmental
Protection Agency (2014)
Rethinking the Components, Coordination, and Management of US Environmental Protection
Agency Laboratories (2014)
Review of the Formaldehyde Assessment in the National Toxicology Program 12th Report
on Carcinogens (2014)
Review of the Styrene Assessment in the National Toxicology Program 12th Report
on Carcinogens (2014)
Review of EPAs Integrated Risk Information System (IRIS) Process (2014)
Review of the Environmental Protection Agencys State-of-the-Science Evaluation of Nonmonotonic
DoseResponse Relationships as They Apply to Endocrine Disruptors (2014)
Assessing Risks to Endangered and Threatened Species from Pesticides (2013)
Science for Environmental Protection: The Road Ahead (2012)
Exposure Science in the 21st Century: A Vision and A Strategy (2012)
A Research Strategy for Environmental, Health, and Safety Aspects of Engineered Nanomaterials (2012)
Macondo WellDeepwater Horizon Blowout: Lessons for Improving Offshore Drilling Safety (2012)
Feasibility of Using Mycoherbicides for Controlling Illicit Drug Crops (2011)
Improving Health in the United States: The Role of Health Impact Assessment (2011)
A Risk-Characterization Framework for Decision-Making at the Food and Drug Administration (2011)
Review of the Environmental Protection Agencys Draft IRIS Assessment of Formaldehyde (2011)
Toxicity-Pathway-Based Risk Assessment: Preparing for Paradigm Change (2010)
The Use of Title 42 Authority at the U.S. Environmental Protection Agency (2010)
Review of the Environmental Protection Agencys Draft IRIS Assessment of Tetrachloroethylene (2010)
Hidden Costs of Energy: Unpriced Consequences of Energy Production and Use (2009)
Contaminated Water Supplies at Camp LejeuneAssessing Potential Health Effects (2009)
Review of the Federal Strategy for Nanotechnology-Related Environmental, Health, and Safety
Research (2009)
Science and Decisions: Advancing Risk Assessment (2009)
Phthalates and Cumulative Risk Assessment: The Tasks Ahead (2008)
Estimating Mortality Risk Reduction and Economic Benefits from Controlling Ozone Air
Pollution (2008)
Respiratory Diseases Research at NIOSH (2008)
Evaluating Research Efficiency in the U.S. Environmental Protection Agency (2008)
Hydrology, Ecology, and Fishes of the Klamath River Basin (2008)
Applications of Toxicogenomic Technologies to Predictive Toxicology and Risk Assessment (2007)
Models in Environmental Regulatory Decision Making (2007)
Toxicity Testing in the Twenty-first Century: A Vision and a Strategy (2007)
Sediment Dredging at Superfund Megasites: Assessing the Effectiveness (2007)
Environmental Impacts of Wind-Energy Projects (2007)
Scientific Review of the Proposed Risk Assessment Bulletin from the Office of Management and
Budget (2007)
Assessing the Human Health Risks of Trichloroethylene: Key Scientific Issues (2006)
New Source Review for Stationary Sources of Air Pollution (2006)
Human Biomonitoring for Environmental Chemicals (2006)
Health Risks from Dioxin and Related Compounds: Evaluation of the EPA Reassessment (2006)
Fluoride in Drinking Water: A Scientific Review of EPAs Standards (2006)
vii
Prepublication copy
State and Federal Standards for Mobile-Source Emissions (2006)

Superfund and Mining MegasitesLessons from the Coeur dAlene River Basin (2005)
Health Implications of Perchlorate Ingestion (2005)
Air Quality Management in the United States (2004)
Endangered and Threatened Species of the Platte River (2004)
Atlantic Salmon in Maine (2004)
Endangered and Threatened Fishes in the Klamath River Basin (2004)
Cumulative Environmental Effects of Alaska North Slope Oil and Gas Development (2003)
Estimating the Public Health Benefits of Proposed Air Pollution Regulations (2002)
Biosolids Applied to Land: Advancing Standards and Practices (2002)
The Airliner Cabin Environment and Health of Passengers and Crew (2002)
Arsenic in Drinking Water: 2001 Update (2001)
Evaluating Vehicle Emissions Inspection and Maintenance Programs (2001)
Compensating for Wetland Losses Under the Clean Water Act (2001)
A Risk-Management Strategy for PCB-Contaminated Sediments (2001)
Acute Exposure Guideline Levels for Selected Airborne Chemicals (nineteen volumes, 2000-2015)
Toxicological Effects of Methylmercury (2000)
Strengthening Science at the U.S. Environmental Protection Agency (2000)
Scientific Frontiers in Developmental Toxicology and Risk Assessment (2000)
Ecological Indicators for the Nation (2000)
Waste Incineration and Public Health (2000)
Hormonally Active Agents in the Environment (1999)
Research Priorities for Airborne Particulate Matter (four volumes, 1998-2004)
The National Research Councils Committee on Toxicology: The First 50 Years (1997)
Carcinogens and Anticarcinogens in the Human Diet (1996)
Upstream: Salmon and Society in the Pacific Northwest (1996)
Science and the Endangered Species Act (1995)
Wetlands: Characteristics and Boundaries (1995)
Biologic Markers (five volumes, 1989-1995)
Science and Judgment in Risk Assessment (1994)
Pesticides in the Diets of Infants and Children (1993)
Dolphins and the Tuna Industry (1992)
Science and the National Parks (1992)
Human Exposure Assessment for Airborne Pollutants (1991)
Rethinking the Ozone Problem in Urban and Regional Air Pollution (1991)
Decline of the Sea Turtles (1990)
Copies of these reports may be ordered from the National Academies Press
(800) 624-6242 or (202) 334-3313
www.nap.edu
viii
Prepublication copy
BOARD ON LIFE SCIENCES

Members
JAMES P. COLLINS (Chair), Arizona State University, Tempe, AZ
ENRIQUETA C. BOND, Burroughs Wellcome Fund, Marshall, VA
ROGER D. CONE, Vanderbilt University Medical Center, Nashville, TN
JOSEPH R. ECKER, Salk Institute for Biological Studies, LaJolla, CA
SEAN EDDY, HHMI Janelia Farm Research Campus, Ashburn, VA
SARAH C.R. ELGIN, Washington University, St. Louis, MO
DAVID R, FRANZ, Former Cdr USAMRIID, Consultant, Frederick, MD
STEPHEN FRIEND, Sage Bionetworks, Seattle, WA
ELIZABETH HEITMAN, Vanderbilt University Medical Center, Nashville, TN
RICHARD A. JOHNSON, Global Helix LLC, Washington, DC
JUDITH KIMBLE, University of Wisconsin, Madison, WI
MARY E. MAXON, Science Philanthropy Alliance, Palo Alto, CA
KAREN E. NELSON, J. Craig Venter Institute, Rockville, MD
ROBERT M. NEREM, Georgia Institute of Technology, Atlanta, GA
MARY E. POWER, University of California, Berkeley, CA
MARGARET RILEY, University of Massachusetts, Amherst, MA
LANA SKIRBOLL, Sanofi, Washington, DC
JANIS C. WEEKS, University of Oregon, Eugene, OR
MARY WOOLLEY, Research!America, Alexandria, VA
Staff
FRANCES E. SHARPLES, Director
JO L. HUSBANDS, Scholar/Senior Project Director
JAY B. LABOV, Senior Scientist/Program Director for Biology Education
LIDA ANESTIDOU, Senior Program Officer, ILAR
KATHERINE W. BOWMAN, Senior Program Officer
MARILEE K. SHELTON-DAVENPORT, Senior Program Officer
KEEGAN SAWYER, Program Officer
AUDREY THEVENON, Associate Program Officer
BETHELHEM M. MEKASHA, Financial Associate
ANGELA KOLESNIKOVA, Administrative Assistant
VANESSA LESTER, Research Associate
KANOKO MAEDA, Senior Program Assistant
JENNA OGILVIE, Senior Program Assistant
ix
Prepublication copy
Preface
The US Department of Defense (DOD) is faced with an overwhelming task in evaluating chemicals
that could potentially pose a threat to its deployed personnel. There are over 84,000 registered chemicals,
and testing them with traditional toxicity-testing methods is not feasible in terms of time or money. In
recent years, there has been a concerted effort to develop new approaches to toxicity testing that incorporate advances in systems biology, toxicogenomics, bioinformatics, and computational toxicology. Given
the advances, DOD asked the National Research Council (NRC) to determine how DOD could use modern approaches for predicting chemical toxicity in its efforts to prevent debilitating, acute exposures to
deployed personnel.
In this report, the Committee on Predictive-Toxicology Approaches for Military Assessments of
Acute Exposures provides an overall conceptual approach that DOD could use to develop a predictivetoxicology system. It reviews the current state of computational and high-throughput approaches for predicting acute toxicity and suggests methods for integrating data and predictions. It concludes with lessons
learned from current high-throughput screening programs and suggests some initial steps for DOD investment.
This report has been reviewed in draft form by persons chosen for their diverse perspectives and
technical expertise in accordance with procedures approved by the NRC Report Review Committee. The
purpose of the independent review is to provide candid and critical comments that will assist the institution in making its published report as sound as possible and to ensure that the report meets institutional
standards of objectivity, evidence, and responsiveness to the study charge. The review comments and
draft manuscript remain confidential to protect the integrity of the deliberative process. We thank the following for their review of this report: Ellen Berg, BioSeek, Inc.; David Clapham, Harvard University;
Mark Cronin, Liverpool John Moores University; Yvonne Dragan, DuPont; John Jenner, Defence Science
and Technology Laboratory; Charles Santerre, Purdue University; Rusty Thomas, US Environmental Protection Agency; Ken Turtletaub, Lawrence Livermore National Laboratory; Daniel Wilson, The Dow
Chemical Company; and Menghang Xia, National Center for Advancing Translational Sciences.
Although the reviewers listed above have provided many constructive comments and suggestions,
they were not asked to endorse the conclusions or recommendations, nor did they see the final draft of the
report before its release. The review of the report was overseen by the review coordinator, David Eaton,
University of Washington, and the review monitor, Mark Cullen, Stanford University. Appointed by the
NRC, they were responsible for making certain that an independent examination of the report was carried
out in accordance with institutional procedures and that all review comments were carefully considered.
Responsibility for the final content of the report rests entirely with the committee and the institution.
The committee gratefully acknowledges the following for their presentations to the committee during open sessions: Alison Director-Myska, Defense Threat Reduction Agency, and Keith Houck, US Environmental Protection Agency.
The committee is grateful for the assistance of the National Research Council staff in preparing this
report. Staff members who contributed to the effort are Ellen Mantus, project director; Marilee SheltonDavenport, senior program officer; Keri Stoever, research associate; James Reisa, director of the Board on
xi
Prepublication copy
Preface
Environmental Studies and Toxicology; Norman Grossblatt, senior editor; Mirsada Karalic-Loncarevic,
manager of the Technical Information Center; Radiah Rose-Crawford, manager of editorial projects; and
Ivory Clarke, senior program assistant.
I especially thank the members of the committee for their efforts throughout the development of this
report.
David Dorman, Chair
Committee on Predictive-Toxicology Approaches
for Military Assessments of Acute Exposures
xii
Prepublication copy
Contents
SUMMARY ......................................................................................................................................................... 3
1
INTRODUCTION........................................................................................................................................ 9
Study Background, 9
The Committee and Its Task, 10
The Committees Approach to Its Task, 11
Organization of This Report, 11
References, 12
CONCEPTUAL FRAMEWORK AND PRIORITIZATION STRATEGY ......................................... 13

Acute Toxicity of Classical Chemical-Warfare Agents, 14
Predicting Acute Toxicity of Potential Chemical-Warfare Agents, 14
A Framework and Strategy for Predicting Acute Toxicity of Potential Chemical-Warfare Agents, 15
Findings and Recommendations, 24
References, 24
NONTESTING APPROACHES RELEVANT TO PREDICTION OF ACUTE

TOXICITY AND POTENCY ................................................................................................................... 28
Initial Chemical Characterization, 28
Use of Physicochemical Properties to Predict Physical Hazards, Chemical Reactivity,
and Pharmacokinetics, 31
In Silico Approaches for Predicting Toxic Effects, 35
Ensuring Scientific Confidence In (Q)SAR Models, 41
References, 43
ASSAYS FOR PREDICTING ACUTE TOXICITY .............................................................................. 50

In Vitro Assays, 51
Nonmammalian In Vivo Animal Models, 58
Emerging Technologies, 61
Metabolic Considerations, 63
Assay Considerations for Improving Prediction of Acute Toxicity, 65
References, 68
INTEGRATION AND DECISION-MAKING FOR PREDICTIVE TOXICOLOGY ........................ 77

General Approach to Integration and Decision-Making, 77
Approaches to Integration, 80
Approaches to Categorization, 87
Findings, 89
References, 89
xiii
Prepublication copy
Contents
LESSONS LEARNED AND NEXT STEPS ............................................................................................ 93

Modern Approaches for the Assessment of Acute Chemical Toxicity, 93
Implementation of the Committees Conceptual Model for Assessment of Acute Chemical Toxicity, 95
Development of a Modern Tiered Approach for Predicting Acute Toxicity: Initial Steps, 96
Other Considerations, 102
References, 105
APPENDIXES
A BIOGRAPHICAL INFORMATION ON THE COMMITTEE ON

PREDICTIVE-TOXICOLOGY APPROACHES FOR MILITARY
ASSESSMENTS OF ACUTE EXPOSURES......................................................................................... 109
B
AVAILABLE DATA OR DATABASES ................................................................................................ 112

BOXES, FIGURES, AND TABLES
BOXES
1-1
2-1
2-2
3-1
3-2
3-3
4-1
5-1
5-2
6-1
6-2
6-3
6-4
6-5
Statement of Task, 10
Definitions of Relevant Terms, 13
Tiered Approach to Predicting Toxicity, 23
Definitions of Selected Nontesting Approaches, 29
Primary Data Considered During a Preliminary Characterization of a Chemical of Interest, 30
In Silico Approaches for Predicting Physicochemical Properties, 30
ACuteTox Testing Strategy, 66
Simplified Illustration of Integration and Decision-making, 78
Example of Meta-Analytic Approach That Uses Irreproducible Discovery Rates to Integrate
Categorization Decisions, 82
An Example of an Integrated Testing Strategy for Predicting Drug-Induced Liver Injury, 95
The Use of Computational Approaches for Evaluating Chemical-Induced Inhibition of
Acetylcholinesterase, 97
The Use of HTS Assays for Identifying Endocrine Disrupting Potential, 99
The Use of the BALB/3T3 Neutral-Red Uptake Cytotoxicity Assay to Predict Acute Toxicity, 100
The Use of HTS Assays to Evaluate Inhibition of Acetylcholinesterase, 103
FIGURES
S-1
S-2
2-1
2-2
3-1
3-2
4-1
Conceptual framework and examples of databases, assays, models and tools for predicting acute
chemical toxicity, 4
Prioritization strategy based on a tiered approach for using predictive-toxicology models and
tools to evaluate agents for acute toxicity, 5
Conceptual framework and examples of databases, assays, models and tools for predicting acute
chemical toxicity, 19
Conceptual framework for the future development of (Q)SARs, 40
Key elements associated with evaluating the adequacy of a (Q)SAR model and its prediction as
adapted from the REACH technical guidance, 42
Intended target families and subfamilies for the ToxCast program, 52
xiv
Prepublication copy
Contents
5-1
5-2
5-3
6-1
Illustration of a general approach to integration and decision-making for applying predictive

approaches to acute, debilitating toxicity, 79
Approaches for integrating disparate data sets with LD50 as an example, 81
ToxPi model for integration of acute-toxicity potential, 84
TABLES
2-1
3-1
3-2
3-3
Biological Processes and Cellular Targets Associated with Acute Toxicity in Humans or
Laboratory Animals, 16
Examples of Heuristic Rules to Predict Oral Absorption, 32
Examples of (Q)SARs for Various Chemical Classes, 36
Examples of Models and Tools to Predict Acute Oral Toxicity, 37
xv
Prepublication copy
APPLICATION OF MODERN TOXICOLOGY

APPROACHES FOR PREDICTING ACUTE
TOXICITY FOR CHEMICAL DEFENSE
Summary
As part of its mission to provide military forces, the US Department of Defense (DOD) must anticipate, defend, and safeguard its personnel against chemical threats. Many factors can determine whether a
chemical agent could pose a threat, and toxicity clearly is one of them. To assess toxicity, DOD has relied
primarily on traditional toxicity testing in which adverse biological responses are measured in laboratory
animals that are exposed to high doses of a test agent. The traditional approaches, however, are expensive
and time-intensive, raise questions about the applicability of results to human populations, raise concerns
about animal welfare, and are impractical for evaluating quickly large numbers of chemicals that could be
used against deployed forces. In recent years, various agencies and organizations have attempted to incorporate advances in systems biology, toxicogenomics, bioinformatics, and computational toxicology to
develop cost-effective approaches for predicting chemical toxicity. Given the recent advances and developments in toxicity-testing methods and approaches, DOD asked the National Research Council (NRC) to
determine the feasibility of developing a toxicity-testing program that uses modern approaches to identify
acutely toxic agents rapidly that are relevant to DOD.1 In response to that request, the NRC convened the
Committee on Predictive-Toxicology Approaches for Military Assessments of Acute Exposure, which
prepared the present report.
CONCEPTUAL FRAMEWORK AND STRATEGY
As requested by DOD, the committee developed an overall conceptual approach that uses modern
approaches for predicting acute, debilitating chemical toxicity. Its approach consisted of three components: (1) a conceptual framework that links chemical structure, physicochemical properties, biochemical
properties, and biological activity to acute toxicity; (2) a suite of databases, assays, models, and tools that
are based on modern in vitro, nonmammalian in vivo, and in silico approaches that are applicable for prediction of acute toxicity; and (3) a tiered prioritization strategy for using databases, assays, models, and
tools to predict acute toxicity in a manner that balances the need for accuracy and timeliness. The committee based its conceptual framework (Figure S-1) on the premise that whole-animal toxicity can be predicted by using information about lower levels of complexity, even down to the level of chemical structure. Specifically, it is hypothesized that chemical structure, physicochemical properties, biochemical
properties, or biological activity in isolated cells and tissues or in nonmammalian organisms can be used
to predict acute mammalian toxicity.
The prioritization strategy was formulated on the basis of DODs stated need to understand the relative threat of the growing list of registered chemical substances. Although the committee cannot prescribe
exactly how to manage various policy tradeoffs, such as the tolerance for false negatives and the
timeframe required for identifying important hazards, it recommends a tiered prioritization strategy (Figure S-2) that applies increasingly complex approaches to place chemicals into three categories: high confidence of low toxicity, high confidence of high toxicity, and uncertain toxicity because of data inadequacy. The first category allows some chemicals to be deselected on the basis of low acute toxicity, and the
emphasis on high confidence indicates a low tolerance for false negatives. The second category allows
chemicals to be selected on the basis of high acute toxicity, and the emphasis on high confidence indi1
The verbatim statement of task is provided in Chapter 1 of this report.
Prepublication Copy
3
cates the need to focus rapidly on chemicals that might pose a risk. The third category represents chemicals that would move to the next tier. Chemicals could be deselected at any stage by considering other
factors, such as chemical availability and weaponizability, that could eliminate them from further consideration. As illustrated in the figure and discussed further in the sections that follow, the testing strategy
proceeds through a number of tiers that are successively more predictive and resource-intensive, from
initial characterization (Tier 0) to nontesting approaches (Tier 1) to high-throughput and mediumthroughput assays (Tier 2) and ultimately to traditional animal testing (Tier 3). Progression through the
tiers requires intermediate integration steps that consider the diversity of data both within a tier and across
tiers. At each tier, DOD will need to develop policies that are relevant to its mission on how to assign
chemicals to various categories and to determine the extent of end-point coverage that is adequate for it to
make reliable decisions. The committee notes that an end point could be a clinical outcome or a molecular
initiating event. If science advances in such a way that adverse outcome pathways of interest to DOD are
known, the strategy shown in Figure S-2 could rely on nontesting and biological assay-based approaches
that evaluate molecular initiating events or measurable key events in the pathways.
NONTESTING APPROACHES FOR PREDICTING ACUTE TOXICITY
The committee envisions that nontesting approaches will be an important component of its conceptual framework. Nontesting approaches range from grouping chemicals that are structurally similar to developing quantitative structureactivity relationship (QSAR) models. The underlying assumption of nontesting approaches is that chemical properties that determine how a chemical will interact with a defined
biological system are inherent in its molecular structure and thus that structurally similar chemicals
should have similar biological activity. The starting point in the application of any nontesting approach is
to search for and evaluate information on the chemical of interest. That step constitutes Tier 0 in the
committees proposed strategy (Figure S-2).
ConceptualFramework
ChemicalStructure,
Physicochemical
Properties
EmpiricalCorrelations
OrganSystemToxicityin
Mammals
Toxicokinetics
BiochemicalProperties,
BiologicalActivityinCells
andLowerOrganisms
MechanisticPathways
WholeOrganismToxicity
inMammals
Databases,Assays,Models,andTools
ToxicityEstimateOutputs
DatabaseandAssayInputs
Chemicalstructure
(e.g.,functionalgroups,molecular
descriptors)
Physicochemical
(e.g.,pH,pKa,KOW)
Biologicalassays
(e.g.,receptorbinding,cytotoxicity,
nonmammalian invivo)
Prioritization
Strategy
ModelsandTools
Readacrosstools
(Q)SARmodelsandtools
Concentrationresponse
models
Toxicokinetic models
Integratedmodels
Mechanismspecific
(e.g.,AC50formitochondrial
dysfunction)
Organsystemspecific
(e.g.,ED50 fornervous,
cardiovascular,respiratory,hepatic,
renal,skeletomuscular,orimmune
system)
Nonspecific
(e.g.,ratLD50,cytotoxicityAC50)
FIGURE S-1 Conceptual framework and examples of databases, assays, models, and tools for predicting acute
chemical toxicity.
Prepublication Copy
Summary
Tier0
Initial
Character
ization
Inadequatedata
+
Nontesting
approaches
Hightoxicity
Lowtoxicity
Lowthreat
duetofactors
unrelatedto
toxicity
Tier1
Integration
andDecision
Making
Hightoxicity
Lowtoxicity
Inadequatedata
+
Biological
assaybased
approaches
Lowthreat
duetofactors
unrelatedto
toxicity
Tier2
Integration
andDecision
Making
Hightoxicity
Lowtoxicity
Inadequatedata
Tier3
Lowthreat
duetofactors
Mammalianinvivotestingor unrelatedto
R&Dtoimprovepredictions
toxicity
FIGURE S-2 Prioritization strategy based on a tiered approach for using predictive-toxicology models and tools to
evaluate agents for toxicity.
As would be expected, information on physical properties, solvation properties, and molecular attributes (physicochemical data) is critical. Physicochemical data can be used to predict a chemicals physical hazard, its reactivity, and its pharmacokinetics, including absorption by different exposure routes,
distribution in the body, and likely metabolites. Physicochemical data can be obtained from the literature,
derived experimentally, or predicted with various in silico techniques. However, many tools that can be
used to predict physicochemical properties have limited chemical applicability; that is, they are most applicable for small organic chemicals.
Nontesting approaches have been used to predict acute toxicity. Specifically, a few (Q)SAR models
have been developed for predicting in vivo acute toxicity.2 Most have focused on the prediction of acute
rodent oral toxicity, such as estimation of oral LD50 values;3 few attempts have been made to derive models for acute toxicity via other exposure routes, such as inhalation and dermal exposure. Nontesting approaches also have been used to predict toxicity end points, such as neurotoxicity or cytotoxicity. More
recent efforts have investigated the integration of in vitro assay data with nontesting approaches to
2
3
The committee uses the shorthand notation (Q)SAR to indicate both SAR and QSAR.
An LD50 value is the dose at which 50% of the population dies.
Prepublication Copy
5
strengthen predictions. Key issues with nontesting (and all other) approaches are their relevance and applicability for the broad array of chemicals of interest to DOD and the reliability and validity of the data
used to develop the models. Furthermore, the exposure routes of interest to DOD are most likely inhalation and dermal exposure, and few nontesting approaches address these exposure routes.
BIOLOGICAL ASSAYS FOR PREDICTING ACUTE TOXICITY
In vitro assays and nonmammalian in vivo assays are important components of the committees
conceptual framework. Numerous screening assays have been developed to measure specific biological
activities. The various assay types are described below with some key limitations noted for DODs purposes.
Specific-Protein Assays. Many enzyme and receptor-binding assays have been developed to examine specific mechanisms of action at the molecular level. Somesuch as ones that measure chemicalinduced inhibition of acetylcholinesterase activity, altered electron transport in mitochondria, and modulation of ion-channel activitymight be relevant for predicting acute toxicity. Although the protein assays hold some promise, a key limitation is that acute toxicity that is not mediated by chemical action on
specific enzymes or receptors will go undetected in these types of assays.
Cell-Based Phenotypic Assays. These assays typically use cultured cells and measure some overall phenotypic output relevant to predicting acute toxicity, such as cellular proliferation, plasma membrane permeability, and adenosine triphosphate content. There is a growing literature on their application
as toxicity screens, especially in drug development. Cell-based assays, particularly ones for evaluating
cytotoxicity, have demonstrated success in predictive toxicology. A key limitation of cytotoxicity assays
is that they do not provide data on some of the most important toxic mechanisms, specifically ones that
involve organ-specific or cell-typespecific physiology. Another limitation of many existing cell-based
assays is that they rely on immortalized cell lines that have little metabolic capability.
Organotypic Models. Organotypic models more closely mimic the anatomy of organs and have
been developed for the skin, eye, lung, liver, and central nervous system. They are especially attractive
given their theoretical potential to model metabolism, biodistribution, and biological activity of a chemical in an in vitro system. However, the science of modeling human organs in a culture dish accurately,
especially in formats suitable for high-throughput testing, and its application to toxicology are still in their
infancy.
Nonmammalian in vivo Assays. In addition to in vitro assays, the committee envisions nonmammalian animal models as a potentially important component of its conceptual framework. Traditional
whole-animal assays have been crucial in understanding how chemicals affect metabolism and exhibit
pathology at the cell and organ level. However, traditional assays are often expensive, require large
amounts of chemicals, and cannot be adapted to even a medium-throughput format. For those and other
reasons, alternative animal models have been developed. Ones that are potentially valuable for adapting to
high-throughput screening rely on the fruit fly (Drosophila melanogaster), a nematode (Caenorhabditis
elegans), and the zebrafish (Danio rerio). One particular advantage of the alternative models is the ability
to identify whole-organism or organ-level responses. However, as with all animal models, a key limitation is related to species differences and use of resulting data to extrapolate to human responses. Furthermore, measuring some end points with alternative animal models has lower throughput than many in vitro
assays, and little is known about their applicability to the assessment of acute toxicity of chemicals that
are relevant to DOD.
In vitro assays, alternative animal models, and other emerging technologies described here and in
more detail later in the committees report hold promise, but some important limitations or considerations
should be noted. First, in vitro assays for predicting acute toxicity have focused primarily on nonmechanistic indicators of toxicity, such as cytotoxicity; they were not developed with a quantitative linkage to
Prepublication Copy
Summary
any phenotype (acute or chronic). Second, existing assays focus on oral exposure; there has been little
consideration of dermal or inhalation exposure. Third, most current in vitro assays do not account for important pharmacokinetic characteristics, such as metabolism, that can influence in vivo toxicity. Fourth,
the nominal chemical concentration used in the assays is not necessarily representative of the concentration at which chemical bioactivity is observed. Fifth, cellular systems commonly use immortalized cancer
cell lines, which might fail to detect chemical activity or effects that might occur in normal (nontumor)
differentiated cells. Sixth, cells can have different levels of activity or responsiveness, depending on
whether they are primary cells, differentiated cells, or immortalized cells and on how many times they
have been cultured, so assay reproducibility can be a problem. Seventh, interpreting activity or effective
concentrations that result from a high-throughput screening assay can be difficult because activity at high
concentrations could represent nonspecific effects and offer little information about specific bioactivity.
Conversely, the absence of activity could mean that the tested concentration is below the in vitro effective
concentration, that the assay does not represent the biological target, or that there are problems with assay
reliability. Current efforts in high-throughput screening support the observations noted here, and the
committee emphasizes that DOD should use the experience from current high-throughput screening programs to design its screening program to predict acute, debilitating toxicity.
INTEGRATION AND DECISION-MAKING FOR PREDICTIVE TOXICOLOGY
A robust integration and decision-making strategy is needed as part of the committees suggested
tiered prioritization strategy (shown in Figure S-2). As noted, the goal of each tier is to place a chemical
into one of three categories: high confidence of high toxicity, high confidence of low toxicity, or inadequate data. That activity will require integrating various data streams and predictions that inform a single
acute-toxicity end point (withinend-point integration and decision-making) and integrating predictions
from several acute-toxicity end points (crossend-point integration and decision-making). The committees report discusses various methods for integrating data and predictions. Key tasks for DOD will be to
define the most informative end points for its purpose (for example, neurotoxicity vs seizures), to set
boundaries or toxicity thresholds for what is considered high or low toxicity for each end point, and
to specify the level of confidence needed to make determinations.
One simple approach for integrating multiple end points is to summarize the categorization results
for each end point in a scorecard. Each end point would be evaluated as to whether the chemical exhibited high toxicity, low toxicity, or inadequate data. A chemical would then be assigned to a high
toxicity overall bin if at least one of the end points scored as high toxicity, a low toxicity overall bin
only if all the end points scored as low toxicity, and an inadequate data overall bin if neither of the
first two conditions is met. That simple approach has the advantage of retaining the end-pointspecific
information to inform future data generation. It is also consistent with a low tolerance for false negatives
in that each end point serves as sufficient evidence to assign a chemical to a high toxicity overall bin.
It is possible to use more complex recombination approaches that would not depend strictly on a
simple decision rule related to the categories for each end point. For example, one approach would be to
provide a summary measure that consisted of a weighted sum of individual toxicity end points. Even if
each individual end point is rated as inadequate data, it is conceivable that the presence of multiple end
points close to their corresponding toxicity thresholds would permit a chemical to be categorized as
high or low on the basis of the summary measure. Setting up appropriate decision rules would be a
key policy question for DOD if it chose to go forward with implementing the committees suggested approach for predicting acute, debilitating toxicity.
LESSONS LEARNED AND NEXT STEPS
Several large-scale initiatives have been evaluating in vitro testing methods for their ability to predict human toxicity, and the committee considered them as it debated the feasibility of a predictive testing
Prepublication Copy
7
program for DOD. The US Environmental Protection Agency (EPA) ToxCast program and the European
ACuteTox program demonstrate that in vitro assays have some value for predicting acute toxicity and
provide evidence that an in vitro screening approach is feasible for evaluating the relative threat of a
chemical as an acute hazard. However, most of the assays developed and validated for high-throughput
screening programs were not developed specifically for acute-toxicity testing and so might be of little use
for identifying chemicals that have the potential to cause acute, debilitating injuries in deployed military
personnel. Lessons learned from those programs, however, could provide a great deal of guidance to
DOD in its designing a system that uses high-throughput screening and predictive models to evaluate
acute toxicity.
On the basis of its review, the committee notes several initial steps that DOD could take to implement the tiered prioritization strategy. First, an investment by DOD in computational and high-throughput
screening could yield benefits in characterizing the toxicity of chemicals on which there are few or no
toxicity data. Computational methods for predicting acute toxicity are seeing steady growth, and highthroughput screening might prove useful in excluding chemicals that have low toxic potential and in identifying toxic chemicals of greater concern for further testing. Second, there are data to suggest that DOD
could use simple cytotoxicity assays to identify chemicals that have low acute-toxicity potential and focus
its attention on chemicals that are more toxic. Additional investment would be required to determine
whether the assays are relevant for identifying highly toxic chemicals that could be used against deployed
troops. Third, the development of targeted mechanistically based assays could provide DOD with a useful
resource for understanding and predicting potential toxicity of chemicals; specifically, having explicit
knowledge of the mechanisms of action that lead to acute systemic toxicity would be valuable in the design and validation of integrated prediction methods. Completing the steps described here might require
DOD to use a variety of reference chemicals, including chemicals of concern, to benchmark the results.
Moreover, completing these steps will be facilitated by selecting well-characterized chemicals that can be
used to evaluate the predictiveness of DODs in vitro assays and approaches against in vivo experimental
results.
The committee anticipates that in the next 310 years any tiered testing approach will not be able to
replace fully the need for targeted mammalian in vivo studies to confirm the toxicity of a chemical of interest. Indeed, the state of the science suggests that development of a predictive acute-toxicity program
will require extensive DOD investment in computational modeling approaches, assay development,
methods for extrapolation of in vitro results to in vivo conditions, and data-integration methods. To begin
the investment, the committee recommends that DOD initiate pilot studies that evaluate chemical classes
of highest concern with well-characterized reference chemicals. The pilot studies would allow DOD to
develop the novel assays and tools needed to predict acute chemical toxicity efficiently and accurately
and to evaluate the rate of false negatives and false positives. The pilot studies could also examine how
generalizable the results of various assays and tools are from one chemical class4 to another. That research would allow DOD to begin to address the size of the chemical space needed to make predictions
about unknown chemicals. The committee emphasizes that DOD could benefit from leveraging its efforts
with other federal activities, such as EPAs ToxCast program. Such collaboration would allow DOD to
complete pilot studies more rapidly and maximize the return on its investment.
In this context, chemical class is used broadly to include structurally related chemicals, chemicals that have different mechanisms of action, and chemicals that have different toxic end points, such as hepatotoxicity and neurotoxicity.
Prepublication Copy
1
Introduction
The mission of the Department of Defense (DOD) is to provide the military forces needed to deter
war and to protect the security of our country (DOD 2014). In support of that mission, DOD must protect the health and capabilities of its personnelmany of whom are deployed overseasby anticipating
and safeguarding against chemical and biological threats. Although many factors, such as availability and
dissemination potential, need to be considered in evaluating a potential threat, chemical toxicity is critical
in determining whether an agent could pose a threat if used by an adversary. Given the numbers of registered chemicals and new chemicals registered each year, evaluating chemical toxicity is especially daunting, particularly in terms of time and money, if one uses traditional toxicity-testing methods. In light of
recent advances in toxicity-testing methods and approaches, DOD would like to determine the feasibility
of developing a high-throughput predictive system that could rapidly identify acutely toxic agents and
threat potentials. Accordingly, DOD asked the National Research Council (NRC) to determine how DOD
could use modern approaches for predicting chemical toxicity in its efforts to prevent debilitating acute
exposures of deployed personnel. In response to that request, NRC convened the Committee on Predictive-Toxicology Approaches for Military Assessments of Acute Exposures, which prepared this report.
STUDY BACKGROUND
Toxicity testing reached a turning point in 2007 with the release of the NRC report Toxicity Testing
in the 21st Century: A Vision and a Strategy. The report set forth a vision for transforming traditional toxicity testing by incorporating advances in systems biology, epigenetics, toxicogenomics, bioinformatics,
and computational toxicology. The new system that was described in the report would be based primarily
on in vitro methods that can be used to evaluate changes in biological processes with cells, cell lines, or
cellular components, preferably of human origin. The motivation for the new system was to accomplish
four important goals: (1) to provide broad coverage of chemicals, chemical mixtures, outcomes, and life
stages, (2) to reduce the cost and time of testing, (3) to use fewer animals and cause minimal suffering in
the animals used, and (4) to develop a more robust scientific basis for assessing health effects of environmental agents (NRC 2007).
On release of the NRC report, several federal agencies embraced the proposed vision. A collaboration that has been informally referred to as Tox21 was formed between the National Toxicology Program
of the National Institute of Environmental Health Sciences, the National Center for Computational Toxicology of the US Environmental Protection Agency (EPA), and the Chemical Genomics Center1 of the
National Institutes of Health; the US Food and Drug Administration joined the collaboration later. The
goal of the collaboration has been to advance the vision proposed in the NRC report. EPA launched ToxCast as a separate activity with the goal of developing cost-effective approaches that use high-throughput
technologies to predict chemical toxicity. The European Registration, Evaluation, Authorization, and Restriction of Chemicals (REACH) regulation encourages companies and other organizations to develop
The Chemical Genomics Center is now part of the National Center for Advancing Translational Sciences.
Prepublication Copy
9
alternative methods that would substitute for traditional methods and has ultimately led to various research initiatives. All those programs and efforts have led to development of new methods and assays for
predicting toxicity.
To protect the armed forces and their ability to serve, DODs Defense Threat Reduction Agency
conducts and sponsors scientific research to predict which chemicals might be used by adversaries as
weapons and the toxicity that could occur if such agents were used. To understand potential toxicity at
various doses, DOD has largely used in vivo toxicity testing in laboratory animals. However, that approach is time-consuming and expensive, must consider species differences in response, and does not enable DOD to keep up with the pace of new chemical registration. To address the challenge of elucidating
the toxicity of more chemicals than can be practically tested in whole-animal assays and to address concerns raised with animal testing, DOD asked the NRC to consider the question of whether the new predictive-toxicology approaches being developed could be used to expedite its evaluation of potential chemical
hazard. Specifically, are the new assays and approaches relevant to DODs interest in acute toxicity? If
not, is there research that would enable DOD to use predictive-toxicology approaches to identify acute
chemical threats?
THE COMMITTEE AND ITS TASK
The committee that was convened as a result of DODs request included experts in toxicology,
computational methods, high-throughput approaches, omics, physiologically based pharmacokinetic
modeling, statistics, model validation, and emergency preparedness (see Appendix A for the committees
biographical information). As noted, the committee was asked to consider the new predictive-toxicology
approaches that have been developed in other fields and to determine whether they could be used to meet
DODs needs. The committees verbatim statement of task is provided in Box 1-1.
BOX 1-1 Statement of Task

An ad hoc committee under the auspices of the National Research Council will consider how the
Department of Defense (DOD) could use modern approaches for predicting chemical toxicity in its efforts to prevent debilitating acute exposures to deployed personnel.
DOD needs to understand the relative threat of the increasingly long list of registered chemical
substances, particularly in terms of potential acute hazard. To help DOD achieve its goal to protect its
deployed personnel, this study will consider modern approaches for predicting toxicity and suggest an
overall conceptual approach for using such information to evaluate acute hazards. The committee will
consider the information provided by predictive-toxicology approaches that is increasingly being generated and used in the environmental health and pharmaceutical sectors to enhance or replace information from traditional, empirical testing of chemical safety in animals. The committee will focus on the
assays and approaches that are being developed by the United States and European agencies (for
example, for the Registration, Evaluation, Authorization, and Restriction of Chemicals (REACH) program, the EPA ToxCast effort, and the NIH/EPA/FDA Tox21 program); these might include computational modeling, structure-activity relationship analysis, analysis of physicochemical characteristics,
read-across techniques, and high-throughput screening and other in vitro assays. Specifically, the
committee will discuss the ability of these approaches to predict acute toxicity at levels relevant to
DOD concerns.
In Phase 1 of this study, the committee will comment on the robustness and the relevance of the
current approaches to meet DOD's needs. If the approaches being developed by other agencies do
not address DOD's concerns about acute toxicity, the committee will broadly describe areas of research that could fill the gaps within the next 5 or 10 years. A second phase of the study, undertaken
at the sponsor's request, will provide more detailed recommendations for a research roadmap.
10
Prepublication Copy
Introduction
THE COMMITTEES APPROACH TO ITS TASK
To address its task, the committee held four meetings. In an open session during the first meeting,
the committee heard presentations from the sponsor on its activities. On the basis of those discussions and
the statement of task, the committee focused its attention on approaches that were considered to be most
relevant for predicting acute debilitating2 or life-threatening3 effects and on the organ systems that were
deemed most likely to be affected. The organ systems of highest concern to DOD included the cardiovascular, respiratory, hepatic, renal, skeletomuscular, immune, and nervous systems, including special senses
(vision and hearing). Each organ system was considered by the committee in its deliberations (see Chapter 2, Table 2-1 for further discussion). During the course of its review, the committee sought representative examples that could illustrate nontesting and assay-based approaches to assess acute chemical toxicity; the examples are provided throughout this report. On the basis of its task, the committee excluded
from consideration traditional toxicity-testing assays (in vivo rodent assays).
In response to its task to predict acute toxicity at levels relevant to DOD concerns, the committee
focused its approach on hazard identification, specifically identifying target organ systems and developing toxicity estimates, such as potency estimates. An approach for predicting acute toxicity that involved
converting toxicity estimates to human exposure estimates, as has been taken with some chemical-warfare
agents (Mioduszewski et al. 2002), was considered beyond the committees charge. Furthermore, the
committee interpreted DODs stated interest in understanding the relative threat of chemicals that could
be used by an adversary against deployed US military personnel to mean prioritizing chemicals in terms
of their potential to cause acute toxicity. Thus, the committee was not focused on predicting human clinical signs or identifying at-risk populations. And, the committee did not set bounds for its proposed strategy because it recognized the need for DOD to develop policies to set toxicity thresholds relevant to its
mission.
During the open session of its first meeting, the committee received a presentation from EPA on the
ToxCast program. The committee considered the efforts of that program that were relevant to predicting
acute toxicity and, more broadly, the technical approaches that might inform development of a DOD
acute-toxicity program. A detailed review of the ToxCast program and its associated assays and methods
was considered beyond the scope of the present report.
ORGANIZATION OF THIS REPORT
The committees report is organized into six chapters and two appendixes. Chapter 2 describes a
conceptual framework and components that would be needed to build an approach based on modern predictive-toxicology methods. Chapter 3 describes the use of nontesting approaches, including quantitative
structureactivity relationships, to predict acute chemical toxicity. Chapter 4 provides a brief review of
medium-throughput and high-throughput assays that can be used to predict acute mammalian toxicity.
Chapter 5 addresses integration of the biological and chemical data into toxicity predictions. Chapter 6
presents important lessons learned from previous predictive acute-toxicity efforts and the committees
overall conclusions. The committee also identifies several steps that DOD could begin to take toward developing high-throughput assays and computational approaches to identify chemicals that have the potential to induce life-threatening acute toxicity in deployed personnel. Appendix A contains biographical information on the committee, and Appendix B discusses available toxicity data or databases that one could
use to find toxicity data.
Acute debilitating effects are defined as ones that cause major irreversible morbidity, such as blindness, loss of
limb function, paralysis, and severe hypoxia.
3
A life-threatening effect is a disease or condition that makes the likelihood of death high unless the exposure is
interrupted.
Prepublication Copy
11
REFERENCES
DOD (US Department of Defense). 2014. About the Department of Defense. [online]. Available: http://www.
defense.gov/about/[accessed March 12, 2015].
Mioduszewski, R., J. Manthei, R. Way, D. Burnett, B. Gaviola, W. Muse, S. Thomson, D. Sommerville, and R.
Crosier. 2002. Interaction of exposure concentration and duration in determining acute toxic effects of sarin
vapor in rats. Toxicol. Sci. 66(2):176-184.
NRC (National Research Council). 2007. Toxicity Testing in the 21st Century: A Vision and a Strategy. Washington, DC: National Academies Press.
12
Prepublication Copy
2
Conceptual Framework and Prioritization Strategy
As discussed in Chapter 1, the committee was asked to consider modern approaches for predicting
acute, debilitating chemical toxicity and to suggest an overall conceptual approach that uses emerging
science to evaluate acute hazards to deployed military personnel. This chapter first discusses current and
future needs for toxicity evaluations of chemical-warfare agents, recognizing the increasing number and
types of chemicals that are potentially available to adversaries. It then describes the conceptual framework and strategy developed by the committee for systematically applying modern approaches to the prediction of acute toxicity. The overall approach, which is illustrated in Figures 2-1 and 2-2, consists of
three components (relevant terms are defined in Box 2-1):
A conceptual framework that links chemical structure, physicochemical properties, biochemical
properties, and biological activity to acute toxicity.
A suite of databases, assays, models, and tools that are based on modern in vitro, nonmammalian
in vivo, and in silico approaches applicable to predicting acute toxicity.
A tiered prioritization strategy for using databases, assays, models, and tools to predict acute toxicity in a manner that balances the need for accuracy and timeliness.
Later chapters in this report provide details of the types of databases, assays, models, and tools that are
available for evaluating acute toxicity, their integration, and next steps that are needed to begin implementing the committees framework and strategy.
BOX 2-1 Definitions of Relevant Terms

An assay is a laboratory system designed to measure a physical, chemical, or biological end point.
A model is a quantitative or qualitative representation of a hypothesis that attempts to explain how different observations are related to one another. In the context of this report, the hypothesis typically
concerns how physical, chemical, or biological data (inputs) can be used to predict biological outcomes of a given exposure (outputs) in a whole animal or human qualitatively or quantitatively.
A tool is an application of a model or set of models, such as in a software package, designed to be routinely used in an applied setting as opposed to a research or development setting.
In vitro approaches include high-throughput screening, other in vitro assays, and more complex systems, such as organotypic cell cultures.
Nonmammalian in vivo approaches include fish, amphibian, nematode, and insect models.
In silico approaches include computational modeling, structureactivity relationship analysis, analysis
of physicochemical characteristics, and read-across techniques (see Chapter 3).
Prepublication Copy
13
ACUTE TOXICITY OF CLASSICAL CHEMICAL-WARFARE AGENTS
Historically, most chemical-warfare agents have belonged to the following chemical classes: nerve
agents (such as sarin and soman), blister or vesicant agents (such as phosgene oxime and sulfur mustards),
blood agents (such as cyanide), and pulmonary agents (such as chlorine and phosgene) (DHHS 2014).
Those agents have been well studied, and a detailed mechanistic understanding that is based on human
data is available for some. For example, the organophosphorus (OP) nerve agents are potent inhibitors of
acetylcholinesterase and result in acute cholinergic effects that occur minutes or hours after exposure.
Knowledge of their mechanisms of toxicity can be useful in the development of therapeutic countermeasures (Sharma et al. in press) and in the development of in vitro tests. For example, in vitro methods have
been developed for the evaluation of cholinesterase inhibition by nerve-gas agents (Worek et al. 2007).
Data on some early, sensitive responses to such agents can support development of acute exposure limits
for the general public. For example, a number of studies indicate that pupil constriction (miosis) is the
most sensitive acute response to human exposure to OP nerve agents (such as sarin), and such end points
have been used as part of the basis of acute exposure limits (NRC 2005).
In vivo testing approaches have been developed and applied to assess the toxicity of chemicalwarfare agents. The vast majority of available toxicity information has come from traditional toxicity
studies in which adverse biological responses were measured in laboratory animals that were exposed to
high doses of a test agent. The acute-toxicity data are often used to provide estimates of the amount of an
agent that would be required to kill 50% of a population of test animals, such as a lethal dose 50% (LD50)
or a lethal concentration 50% (LC50). In addition, pharmacokinetic studies that were designed to identify
species differences in chemical absorption, distribution, metabolism, and excretion (Tenberken et al.
2010; Benson et al. 2011a,b) and specialized pharmacokinetic models (such as ones that use a human or
porcine skin flap) that were developed to evaluate absorption and toxicity of some chemical-warfare
agents (Riviere et al. 1995; Monteiro-Riviere and Inman 1997; Vallet et al. 2008) have been undergoing
incremental refinement since their inception. A limitation of the in vivo studies, however, is that they tend
to be low-throughput, require consideration of species differences in response, and often provide little
insight into a chemicals mechanism of action.
PREDICTING ACUTE TOXICITY OF POTENTIAL CHEMICAL-WARFARE AGENTS
Only a few chemicals have been formally classified as chemical-warfare agents. However, the list of
chemicals that could potentially be used by an adversary against deployed US personnel is large and continues to grow as more chemicals enter the marketplace. Therefore, Department of Defense (DOD) efforts
to evaluate potential chemical-warfare agents need to consider a wide array of chemicals beyond traditional chemical-warfare agents, including toxins of biological origin (such as trichothecenes, saxitoxin,
and tetrodotoxin), industrial chemicals (such as ammonia), pesticides (such as sodium monofluoroacetate), and pharmaceutical agents (such as cocaine and amphetamine) (Holstege et al. 2007). The ability of
an adversary to use those or other chemicals will depend on their or their precursors availability and
weaponizability and on other factors that were deemed beyond the scope of the committees work but that
might be important in deciding which agents to evaluate for acute toxicity.
To determine the best way to assess the growing list of registered chemical substances, the committee considered the adverse effects of highly toxic agents, including those of classical chemical-warfare
agents, and identified the following organ systems to be of greatest importance for evaluating acute, debilitating hazards: cardiovascular, respiratory, hepatic, renal, skeletomuscular, immune, and nervous systems, including special senses (vision and hearing). Sufficient perturbation in those organ systems can
lead to a progression in the severity of effects that can result in incapacitation or death of the whole organism.
14
Prepublication Copy

Given ethical considerations, additional acute human-toxicity data are unlikely to be available except in cases of accidental release or deliberate attack for which exposure estimates are typically highly
uncertain or unknown. And, available traditional toxicity-testing data provide little information about
acute, debilitating toxicity. For example, information about chronic, reproductive, or developmental hazardsalthough important for chemical risk assessment in occupational or environmental settingsis of
secondary concern in a military environment where acute, debilitating hazards are of immediate importance. As with other toxicity-testing programs, DOD recognizes that it would be prohibitively expensive and time-consuming to test all potential agents with traditional whole-animal toxicity-testing approaches even if such testing were limited to evaluations of acute toxicity. Moreover, traditional in vivo
testing, particularly for acute toxicity, often does not provide information on the cellular or biological
mechanisms of toxicity or in some cases even identify the target organ system.
Although some of the more modern, biological assay-based approaches have been used to elucidate
mechanisms of action of many of the classical chemical-warfare agents described above, they have not
been used to identify potential chemical-warfare agents. Nonetheless, the fact that some high-throughput
screening data on chemical-warfare agents already exist suggests the feasibility of using such approaches
to evaluate agents and provides important reference data with which results on other agents can be
compared. The modern predictive approaches can also inform decisions as to whether additional mammalian in vivo testing of an agent is needed and might be able to provide information about the cellular and
biological mechanistic events associated with acute toxicity and indicate whether additional testing should
focus on a specific organ system or biological target.
A FRAMEWORK AND STRATEGY FOR PREDICTING ACUTE
TOXICITY OF POTENTIAL CHEMICAL-WARFARE AGENTS
Conceptual Framework
A predictive-toxicology program to assess acute toxicity ideally will build on knowledge about the
cellular targets and mechanisms of action that are related to acute human toxicity. Acute toxicity depends
on fewer biological and chemical pathways than those envisioned by NRC (2007) for a general toxicity
evaluation. It could be more straightforward, although still challenging, to predict the potential for acute
toxicity than the potential for toxicity in the general public in a variety of organ systems, life stages, populations, and exposure timeframes. Specifically, clinical toxicologists have recognized several cellular or
biological targets that are often associated with the acute lethal or debilitating effects of chemicals. Table
2-1 provides an overview of those cellular targets and relevant examples and lists some chemicals that
affect the targets. It should be noted that there is not necessarily a one-to-one correspondence between
mechanistic targets and organ-system targets because multiple mechanisms could affect a single organ
system, a single mechanism could affect multiple organ systems, and debilitation or death could occur
from multiorgan failure.
The relatively detailed knowledge of the multiple mechanisms by which chemicals can cause acute
toxicity supports the basic premise of predictive toxicology that whole-animal toxicity can be predicted
on the basis of information on lower levels of complexity down to the level of chemical structure. That
premise forms the basis of the conceptual framework developed by the committee, illustrated in Figure
2-1. Specifically, it is hypothesized that chemical structure, physicochemical properties, biochemical
properties, or biological activity in isolated cells and tissues or in nonmammalian organisms can predict
acute mammalian toxicity. The predictions can arise through observations of empirical or statistical correlations or through knowledge of the relevant mechanistic pathways, either of which could potentially be
coupled with toxicokinetic information.
Prepublication Copy
15
16
TABLE 2-1 Biological Processes and Cellular Targets Associated with Acute Toxicity in Humans or Laboratory Animalsa, b
Biological Process or Cellular Target
Example Target
Organ System
Examples of in vitro Assay Approachesc
Example
Chemical or Biological Agent
Altered axonal transport
Disruption of microtubule
function
Vinca alkaloids
, '-iminodipropionitrile
Nervous
Tubulin polymerization assessed with flow

cytometry (Morrison and Hergenrother 2012)
Altered impulse conduction by axonal

membrane
Blocking of Na+ ion channel
Tetrodotoxin
Nervous
Cell-based assays of the membrane potential that

use fluorescent dye (Hill et al. 2014)
Reduced precursor availability or

neurotransmitter synthesis and storage
Inhibition of acetylcholine uptake Vesamicol

into synaptic vesicle
Reserpine (dopamine)
Nervous
PC12 cell-based microelectrode assay (Cui et al.

2006; Chen et al. 2008)
Altered neurotransmitter release
Blocking of release of
acetylcholine at neuromuscular
junction
Botulinum toxin
Nervous
PC12 cell-based system for in vitro measurements

of neurotransmitter release events (Yakushenko et
al. 2013)
Presynaptic release of
acetylcholine and other
neurotransmitters
-latrotoxin
Neurotransmitter agonists
Opioids, benzodiazepines, nicotine,

anatoxin-a, kainic acid
Nervous
Neurotransmitter antagonists
Curare, -bungarotoxin,
3-quinuclidinyl benzilate
Review of selected methods to assess receptor

binding (Dunlop et al. 2007); use of stably
transfected HEK cells expressing human D2, D3,
or D4 dopamine receptors as a screening tool
(Vangveravong et al. 2006; Xiao et al. 2014)
Acetylcholinesterase inhibition
Altered dopamine transporter
Altered serotonin reuptake
Altered dopamine reuptake
Nerve gas agents

Cocaine
Fluoxetine
Amphetamine
Nervous
Zebrafish-based (Jin et al. 2013) and enzymebased (Wille et al. 2010) assays for
acetylcholinesterase inhibitors
Altered electrical conduction of heart or

cardiomyocyte contractility
Sodiumpotassium ATPase
blockers
Digoxin
Cardiovascular
Assessment of altered cardiomyocyte contraction

(Himmel 2013; Pointon et al. 2013, 2015; Sirenko
et al. 2013; Scott et al. 2014) and
electrophysiology (Lopez-Izquierdo et al. 2014);
organotypic zebrafish heart model (Pieperhoff et
al. 2014)
Altered ion pump

(Na+, Ca++, K+) activity
Inhibit K+ channel function

Inhibit Na+ channel function
Dendrodotoxin, 4-aminopyridine
Tetrodotoxin, saxitoxin
Cardiovascular
Comparison of in vitro potency of saxitoxin in

cultured neurons with in vivo results (Jellett et al.
1992; Vale et al. 2008).
Change in neurotransmitter function
Altered neurotransmitter binding at

receptor sites
Impaired neurotransmitter
inactivation mechanisms
Altered ion flow
Pore formation
Na+/H+ antiporter
Ionophores
Cardiovascular
Assessment of cell permeability and other end

points in multiple strains of mouse embryonic
fibroblasts (Suzuki et al. 2014)
Ion-channel interactions
Transient receptor potential

cation channel, subfamily A,
member 1 (TRPA1) activation
Sulfur mustard
Acrolein
Respiratory
Role of TRPA1 as a chemosensor (Bch et al.

2013; Stenger et al. in press)
Prepublication Copy
Increased permeability of cellular membranes
Acylation of proteins and lipids

(pulmonary edema)
Phosgene
Respiratory
Human epithelial lung cells as a system to

investigate pulmonary edema (Wijte et al. 2011)
Mitochondrial dysfunction
Multiple mechanisms
Various
Multiple
Various HTS of mitochondrial dysfunction (Jensen

and Rekling 2010; Sakamuru et al. 2012; Vongs et
al. 2011; Attene-Ramos et al. 2013, 2015; Sirenko
et al. 2014b; Wills et al. 2013)
Reduced ATP production
Inhibition of oxidative
phosphorylation
Fluoroacetate, cyanide, chlordecone,

bromethalin
Nervous, cardiovascular,
multiple
Monitoring of ATP production or cell

concentrations (Steinhoff et al. 2015)
Activation of apoptotic pathways
Multiple
Cisplatin, doxorubicin
Multiple
Cell-imaging methods for cultured cardiomyocytes

(Mioulane et al. 2012)
Competitive binding to hemoglobin
Carboxyhemoglobin production
Carbon monoxide
Multiple
In vitro assessment of carbon monoxide and

cyanide binding to hemoglobin using human
blood (Thoren et al. 2013)
Irritant or cytotoxic effects
Pulmonary edema
Phosgene, chlorine, methylisocyanate
Respiratory
Microfluidic system that mimics alveolarcapillary interface of human lung (Huh et

al. 2012)
Lipid peroxidation
Hepatic injury
Acetaminophen, carbon tetrachloride
Hepatic
Lipid peroxidation cell-based and cell-free assays

(Kelesidis et al. 2014)
ROS formation
Renal injury
Aminoglycosides
Renal
HTS assays to measure ROS formation (Adams et

al. 2013; Prasad et al. 2013; Zielonka et al. 2014)
Altered prostaglandin synthesis
Vascular dysfunction
NSAIDs
Cardiovascular
HTS assay for prostaglandin E synthase activity

(Andersson et al. 2012)
Genetic damage
Multiple
Multiple
Multiple
HTS assays to measure genetic damage (Gutzkow

et al. 2013; Li et al. 2013; Wasalathanthri et al.
2013; Watson et al. 2014; Bandi et al. 2014; Falk et
al. 2014; van der Linden et al. 2014)
DNA or protein adduct formation
DNA alkylation
Aflatoxins, cisplatinin (kidney),

sulfur mustard
Multiple
Medium-throughput methods for quantification of

sulfur mustard adducts to proteins (Andacht et al.
2014; Pantazides et al. 2015)
Altered protein synthesis
Inactivation of ribosomes
Ricin
Multiple
Assay for the measurement of adenine released

from ribosomes or small stem-loop RNAs by
ricin toxin A-chain catalysis (Sturm and Schramm
2009)
Disruption of cytoskeleton
Actin or cytoskeleton
disassembly
Phalloidin, microcystin
Multiple
Assessment of cytoskeleton integrity in a

hepatocyte model (Sirenko et al. 2014a)
Altered bioenergetics
Prepublication Copy
Chemical reactivity
Altered oxygen transport
Oxidative stress or ROS formation
Damage to DNA and subcellular systems
(Continued)
17
18
TABLE 2-1 Continued

Example
Chemical or Biological Agent
Example Target
Organ System
Immunogenic interactions with cell

macromolecules
Alteration of mammalian
immune system function
Endotoxin, anthrax exotoxins
Immune
Reviews of endotoxin and anthrax toxins

(Thorn 2001; Liu et al. 2013).
Autoimmunity
Autoimmune hepatitis and

necrotizing myositis
Statins
Multiple
Reviews of statins and myositis (Jones et

al. 2014) and hepatotoxicity (deLemos et
al. 2014)
Biological Process or Cellular Target
Examples of in vitro Assay Approachesc
Immune-mediated effects
The lists of chemicals and biological targets shown here are not intended to be complete; rather, this table shows a variety of plausible biological targets and responses that need to
be considered in evaluating chemicals for acute toxicity that could debilitate or kill deployed troops.
b
Boldface: Listed in Chemical Weapons Convention or is a suspected chemical agent of concern.
c
There are relatively few applications of these methods to the prediction of acute toxicity; thus, the information is provided for illustrative purposes only to demonstrate the types of
approaches used to date (see Chapter 4 for additional information).
Abbreviations: ATP, adenosine triphosphate; DNA, deoxyribonucleic acid; HTS, high-throughput screening; NSAID, nonsteroidal anti-inflammatory drug; PC, pheochromocytoma; RNA, ribonucleic acid; ROS, reactive oxygen species.
Prepublication Copy
ConceptualFramework
ChemicalStructure,
Physicochemical
Properties
EmpiricalCorrelations
OrganSystemToxicityin
Mammals
Toxicokinetics
BiologicalActivityinCells
andLowerOrganisms
MechanisticPathways
WholeOrganismToxicity
inMammals
Databases,Assays,Models,andTools
(Chapters35)
ToxicityEstimateOutputs
DatabaseandAssayInputs
Chemicalstructure
(e.g.,functionalgroups,molecular
descriptors)
Physicochemical
(e.g.,pH,pKa,KOW)
Biologicalassays
(e.g.,receptorbinding,cytotoxicity,
nonmammalian invivo)
Prioritization
Strategy
ModelsandTools
Readacrosstools
(Q)SARmodelsandtools
Concentrationresponse
models
Toxicokinetic models
Integratedmodels
(Fig.22)
Mechanismspecific
(e.g.,AC50 formitochondrial
dysfunction)
Organsystemspecific
(e.g.,ED50 fornervous,
cardiovascular,respiratory,hepatic,
renal,skeletomuscular,orimmune
system)
Nonspecific
(e.g.,ratLD50,cytotoxicityAC50)
FIGURE 2-1 Conceptual framework and examples of databases, assays, models, and tools for predicting acute
chemical toxicity.
Databases, Assays, Models, and Tools

Evaluating the potential for acute toxicity by using the conceptual framework of predictive toxicology requires a suite of databases, assays, models, and tools to cover the relevant physical, chemical, biological, and toxicological space. In general, input information on chemical structure, physicochemical
properties, biochemical properties, and biological activity that is used to make predictions will be obtained from relevant databases or assays. Chemical-structure data might range from chemical-grouping
data (for example, reaction chemistry domains, such as Michael acceptors) to quantitative descriptors of
chemical structure (for example, topological descriptors and semiempirical quantum chemical descriptors). Physicochemical-property data include quantities measured in physical or chemical assays,
such as boiling point, pH, pKa, and KOW.1 Biochemical measures are usually measures of specific molecular interactions (such as DNA binding and receptor activation) with biological molecules, such as nucleic acids, proteins (including enzymes and receptors), and lipids. Finally, biological activity might include both specific measures of function (such as acetylcholinesterase inhibition) and nonspecific
measures of toxicity (such as cytotoxicity from in vitro assays and LC50 estimates obtained from assays
that use Drosophila). Databases and assays for chemical structures and physicochemical properties are
discussed in Chapter 3 and Appendix B, and assays for biochemical properties and biological activity in
Chapter 4.
In this same context, the prediction outputs consist of estimates of end points related to acute toxicity. The end points might be related to particular mechanisms known to cause acute toxicity (see, for
example, Table 2-1), end points related to specific organ system targets (noted above), or nonspecific end
points, such as death and cytotoxicity. Data on those end points for chemicals of known toxicity (such as
classical chemical-warfare agents) can serve as training and test data for building models or tools to
predict the same end points for chemicals on which such data are lacking. Finally, the specific form of the
1
Kow is the octanol-water partition coefficient.
Prepublication Copy
19
outputs might be qualitative (such as active or inactive), semiquantitative (such as a ranking), or quantitative (such as a numerical estimate of dose). However, as discussed further below, quantitative estimates
are likely to be of greatest use for military applications.
A variety of models and tools might be used to provide toxicity estimates. Models and tools might
be qualitative (such as decision trees) or quantitative (such as statistical regression) and might include
statistically based (or machine-learningbased) models, biologically based models, or a mixture of the
two. No model or tool is universally applicable, so it is important that a models or tools domain of applicability is characterized in terms of the chemical space in which it is predictive and the relevant toxicological end points that are covered. Moreover, toxicokinetic models might need to be integrated into the
predictions to address absorption, distribution, metabolism, and excretion relevant to acute toxicity. Finally, models and tools differ with respect to the uncertainty or confidence in their predicted outputs.
As described further in Chapters 3-5, there are many available databases, assays, models, and tools
that could be used to predict acute toxicity. Because they vary in their required level of effort, their relevance to acute toxicity, their domain of applicability, the extent to which they address toxicokinetics, and
the uncertainty or confidence in their predictions, the committee developed an overall strategy for using
them to evaluate acute toxicity. The committees strategy is described next.
Prioritization Strategy for Evaluating Acute Toxicity
Effective implementation of predictive models and tools depends on first identifying the ultimate
(and acceptable) use of the predictive outputs. The committees task states that DOD needs to understand
the relative threat of the increasingly long list of registered chemical substances, particularly in terms of
potential acute hazard. The committee interprets that statement to mean that the goal of the predictivetoxicology approach is to prioritize substances in the sense of identifying those of greater and less concern for acute toxicity. Three key issues must be considered in developing a strategy for prioritization: the
need for quantitative measures of potency and their uncertainties, the need to minimize false negatives,
and the need to screen a large number of chemicals rapidly.
The first key issue is that prioritization with respect to toxicity inherently requires a quantitative measure of potency and a characterization of uncertainty. Ideally, potency should be defined in absolute units,
such as an acute oral LD50 in milligrams per kilogram per day. Relative potency measures might be informative if they include reference chemicals that have known toxicity and, in that case, could be converted to
absolute potency measures if toxicokinetic information is also available to make any necessary adjustments.
Qualitative outputs, such as binary categorizations of active or inactive, might be useful as an additional
output to target testing for specific end points but are not useful by themselves. Furthermore, in the absence
of human data, there will always be inaccuracies in predicting human toxicity, so it is important to characterize the uncertainty or confidence associated with any predicted potency value. Because a decision-maker
might have defined tolerance for errors (such as for false negatives and false positives2), the degree of uncertainty or confidence in a prediction can influence the decision that is made about a particular substance.
Therefore, an estimated confidence interval is essential to any prioritization strategy.
The topic of uncertainty leads to the second key issue: given that this task is meant to prevent death
and debilitating injuries of US military personnel, it is expected that there will be a low tolerance for false
negatives. A likely consequence of reducing the number of false negatives is that a higher percentage of
chemicals will be retained for assessment with more accurate but more resource-intensive approaches. The
overall time needed to complete the review for the whole chemical space would increase accordingly. However, the timeframe to complete an assessment of thousands of chemicals could be unacceptably long, and a
chemical could be successfully weaponized in that timeframe before a decision has been made.
2
In this context, a false negative occurs when a chemical is identified as having low toxicity when it actually
has high toxicity, and false positive occurs when a chemical is identified as having high toxicity when it actually
has low toxicity.
20
Prepublication Copy

The timeframe raises the third key issue: the prioritization strategy needs to be able to screen chemicals in a manner that allows rapid identification of the ones that pose the greatest risk. A rapid-screening
scenario could be acceptable if follow-up screening is conducted to ensure that all potential chemical
threats are eventually identified. It is critical that such an approach incorporate a short timeline that progresses efficiently through a multitiered approach to allow timely reconsideration of chemicals that are
not originally classified as posing the greatest risk. Lessons learned from the first round of screening
could then be leveraged effectively in the reassessment and enable a more informed review and follow-up
validation of the initial approach that can also be rapidly implemented. The risk of using this approach
lies in a time lag that could result in weaponization of a chemical that was originally not deemed to pose a
great threat.
The policy tradeoff of balancing a low tolerance for false negatives with a need to identify important
hazards rapidly is beyond the scope of the committees charge. However, as a general approach, the
committee found that the policy tradeoff could be managed through a tiered prioritization approach as
illustrated in Figure 2-2. Specifically, the committees proposed prioritization strategy proceeds through a
number of tiers that apply successively more predictive and resource-intensive approaches than the previous ones. At each tier, a chemical is placed into one of three general categories:
(a) High confidence of low toxicity. These chemicals would be deselected for further study and are
considered to have a low relative acute toxicity. The requirement that the determination be made with
high confidence addresses the low tolerance for false negatives.
(b) High confidence of high toxicity. These chemicals would be selected and considered to have a
high relative acute toxicity. The requirement that the determination be made with high confidence focuses
attention quickly on chemicals that might pose a high risk.
(c) Uncertain toxicity due to inadequate data. The remaining chemicals would be candidates for
moving to the next tier of evaluation for acute toxicity.3 The uncertainties might stem from available predictions of high uncertainty or low confidence or from inadequate coverage of end points deemed important for evaluating acute toxicity. Depending on resource constraints, it might be reasonable to assess
the chemicals by using additional factors unrelated to toxicity, such as weaponizability. Thus, some
chemicals might be further deselected for further study because they pose a low threat owing to factors
unrelated to toxicity (discussion of such factors is beyond the committees charge). If additional evaluation of toxicity is determined to be needed, the chemical would be moved to the next tier of hazard evaluation to reduce uncertainty concerning the potential for acute, debilitating toxicity. Uncertainty might also
be reduced through additional research into and development of approaches to improve acute-toxicity
prediction, that is, by decreasing the number of chemicals in category (c) and increasing the ability to discriminate between categories (a) and (b).
The categorization can be based on a single end point (possibly based on multiple approaches) or on multiple end points. The committee notes that an end point could be a clinical outcome or a molecular initiating event (see Figure 3-1). If science advances in such a way that adverse-outcome pathways of interest to
DOD are known, the strategy shown in Figure 2-2 could rely on nontesting and biological assay-based
approaches that evaluate molecular initiating events or measurable key events in the pathways.
The committee broadly grouped the available approaches to predicting acute toxicity into four tiers,
beginning with an initial chemical characterization (Tier 0), proceeding to nontesting approaches (Tier 1),
then to biological assay-based approaches (Tier 2), which includes nonmammalian animal species, and
ultimately to traditional whole-animal toxicity testing (Tier 3). The tiers are described further in Box 2-2.
Some chemicals in categories (a) and (b) might be carried to a higher tier for validation purposes.
Prepublication Copy
21
Tier0
Initial
Character
ization
(Chapter3)
Inadequatedata
+
Nontesting
approaches
(Chapter3)
Hightoxicity
Lowtoxicity
Lowthreat
duetofactors
unrelatedto
toxicity
Tier1
Integration
andDecision
Making
(Chapter5)
Inadequatedata
+
Biological
assaybased
approaches
(Chapter4)
Hightoxicity
Lowtoxicity
Lowthreat
duetofactors
unrelatedto
toxicity
Tier2
Integration
andDecision
Making
(Chapter5)
Hightoxicity
Lowtoxicity
Inadequatedata
Tier3
Lowthreat
duetofactors
toxicity
FIGURE 2-2 Prioritization strategy based on a tiered approach for using predictive-toxicology models and tools to
evaluate agents for acute toxicity. The strategy can be applied to a single end point (such as lethality, neurotoxicity,
and cytotoxicity) and to multiple end points.
A key step in each tier is integration and decision-making (described in Chapter 5). Even within a tier, such as nontesting approaches (Tier 1), there might be diverse outputs and predictions from different
models or tools that need to be synthesized. For example, a simple integration approach could be in the
form of a scorecard that counts positive and negative results from available nontesting approaches; a
more sophisticated integration approach might aggregate different predictions. In addition, Tier 2 integration should consider the previous results of nontesting approaches with the newly generated biological
assay data. Absorption, distribution, metabolism, and excretion (ADME) considerations can be integrated
to provide relevant information, such as chemical bioavailability or distribution to target organs.
22
Prepublication Copy

The committee envisions that a decision as to whether a chemical is categorized as having high toxicity, low toxicity, or inadequate data could be made for each end point that is relevant to acute toxicity
(see examples in Figure 2-1). As discussed previously, such decisions would be based on quantitative toxicity estimates for each toxicity end point and associated levels of confidence or confidence intervals. Defining the specific thresholds for assigning a chemical to each category will require expert judgment on
the part of DOD. However, reference chemicals with known high and low toxicities could help to inform
those boundaries. Overall, chemicals would also be assigned to categories for multiple individual end
points that reflect different types of acute toxicity although, as noted in Chapters 3 and 4, there are many
gaps in coverage of end points related to acute toxicity at all tiers. Therefore, noting the gaps as part of the
prioritization strategy provides guidance on how to target testing in later tiers. And, it is up to DOD to
determine the extent of coverage of end points that is adequate for it to make sufficiently reliable decisions at each tier.
BOX 2-2 Tiered Approach to Predicting Toxicity

A tiered approach to predicting toxicity consists of successively more predictive and resourceintensive approaches to evaluating toxicity (see Figure 2-2). DOD might deselect a chemical at any tier
on the basis of factors unrelated to toxicity, such as availability or weaponizability.
Tier 0 would be an initial chemical characterization of toxicity and physicochemical properties based
on existing data. In addition to characterizing acute toxicity, traditional toxicity data can be used to
build and test predictive-toxicology models in Tiers 1 and 2, and physicochemical data might be important for understanding potential exposure routes, bioavailability, target-tissue distribution, and potential physical hazards or chemical reactivity associated with an agent. Chapter 3 and Appendix B
discuss the availability, accessibility, and sources of acute-toxicity data and other data useful for initial
chemical characterization.
Tier 1 uses models and tools that make predictions based on chemical structure and physicochemical
properties. Such models and tools, discussed in Chapter 3, are termed nontesting approaches because they do not involve any additional toxicity testing and data generation. Such approaches include
the use of structureactivity relationships, quantitative structureactivity relationships, and readacross. As discussed further in Chapter 3, the available approaches and tools differ in their potential
applicability to prediction of acute toxicity, their chemical domain of applicability, and their predictive
power and degree of uncertainty. There are a number of gaps in chemical space, biological space,
and predictivity; for many chemicals or end points, predictions based on nontesting approaches will
often be highly uncertain.
Tier 2 is the conduct of biological assays to generate data to reduce uncertainty in the toxicity evaluation. Biological assays in this tier include specific protein assays, cell-based phenotypic assays,
organotypic models, and nonmammalian in vivo animal models. Toxicity predictions based on such
data, discussed in Chapter 4, are termed biological assay-based. Ideally, this biological testing focuses on specific biological targets that are based on information from previous tiers. However, it is also
likely to include nonspecific toxicity end points, such as cytotoxicity. As with nontesting approaches,
available biological assay-based approaches and tools differ in their potential applicability to prediction
of acute toxicity, their chemical domain of applicability, and their predictive power and degree of uncertainty. There are a number of gaps in chemical space, biological space, and predictivity; for many
chemicals or end points (although one hopes fewer than in Tier 1), predictions based on biological
assay-based approaches will be highly uncertain.
Tier 3 is the conduct of mammalian in vivo testing. These traditional approaches are not part of the
committees task. However, the committee notes that as in Tier 2, ideally this toxicity testing will focus
on specific biological targets that are based on information from previous tiers. There could also be
specific gaps or limitations identified in earlier tiers that could be addressed with additional research or
development of new models and tools.
Prepublication Copy
23
FINDINGS AND RECOMMENDATIONS
Finding: There are multiple mechanisms by which chemicals can elicit acute, debilitating toxicity,
and these mechanisms provide support for a predictive-toxicology conceptual framework that predicts
system, tissue, or organism toxicity on the basis of chemical structure, physicochemical properties, biochemical properties, or biological activity in isolated cells, tissues, and lower organisms.
Finding: Such a conceptual framework that includes databases, assays, models, and tools that are
applicable to prediction of acute toxicity could be used to evaluate a large number of chemicals for acutetoxicity potential more rapidly than traditional, mammalian in vivo studies.
Finding: In prioritizing chemicals in terms of their potential to cause acute toxicity, DOD will
need to balance a relatively low tolerance for false negatives with a need to evaluate a large number of
chemicals rapidly. Regardless of how DOD decides to balance those objectives, they can be managed
through a tiered prioritization strategy that applies successively more predictive and resource-intensive
approaches as needed.
Recommendation: The committee recommends a prioritization strategy that broadly groups approaches to prediction of acute toxicity into four tiers, beginning with an initial chemical characterization
(Tier 0), moving to nontesting approaches (Tier 1), then to biological assay-based approaches (Tier 2),
and finally to traditional mammalian in vivo testing (Tier 3). Progression through the tiers will require
intermediate integration steps that consider the diverse data within a tier and among tiers. The prioritization strategy can be applied to single or multiple end points.
Recommendation: As part of the prioritization strategy, the committee recommends placing
chemicals into one of three general categories at each tier: high confidence of high toxicity, high confidence of low toxicity, and inadequate data to evaluate toxicity confidently. Chemicals placed in the
last category, inadequate data, are moved to the next tier for additional, more resource-intensive evaluation. Quantitative estimates of how potent the chemicals might be and of the confidence or uncertainty in
each estimate will be needed to place chemicals into categories. DOD will need to use expert judgment to
define specifically how chemicals are to be assigned to the different categories.
REFERENCES
Adams, D.J., Z.V. Boskovic, J.R. Theriault, A.J. Wang, A.M. Stern, B.K. Wagner, A.F. Shamji, and S.L. Schreiber.
2013. Discovery of small-molecule enhancers of reactive oxygen species that are nontoxic or cause genotypeselective cell death. ACS Chem Biol. 8(5):923-929.
Andacht, T.M., B.G. Pantazides, B.S. Crow, A. Fidder, D. Noort, J.D. Thomas, T.A. Blake, and R.C. Johnson. 2014.
An enhanced throughput method for quantification of sulfur mustard adducts to human serum albumin via isotope dilution tandem mass spectrometry. J. Anal. Toxicol. 38(1):8-15.
Andersson, S., M. Norman, R. Olsson, R. Smith, G. Liu, and J. Nord. 2012. High-precision, room temperature
screening assay for inhibitors of microsomal prostaglandin E synthase-1. J. Biomol. Screen. 17(10):13721378.
Attene-Ramos, M.S., R. Huang, S. Sakamuru, K.L. Witt, G.C. Beeson, L. Shou, R.G. Schnellmann, C.C. Beeson,
R.R. Tice, C.P. Austin, and M. Xia. 2013. Systematic study of mitochondrial toxicity of environmental chemicals using quantitative high throughput screening. Chem. Res. Toxicol. 26(9):1323-1332.
Attene-Ramos, M.S., R. Huang, S. Michael, K.L. Witt, A. Richard, R.R. Tice, A. Simeonov, C.P. Austin, and M.
Xia. 2015. Profiling of the Tox21 chemical collection for mitochondrial function to identify compounds that
acutely decrease mitochondrial membrane potential. Environ. Health Perspect. 123(1):49-56.
Bandi, S., P. Viswanathan, and S. Gupta. 2014. Evaluation of cytotoxicity and DNA damage response with analysis
of intracellular ATM signaling pathways. Assay Drug Dev. Technol. 12(5):272-281.
Benson, J.M., B.M. Tibbetts, W.M. Weber, and G.R. Grotendorst. 2011a. Uptake, tissue distribution, and excretion
of 14C-sulfur mustard vapor following inhalation in F344 rats and cutaneous exposure in hairless guinea pigs.
J. Toxicol. Environ. Health A 74(13):875-885.
Benson, J.M., A.P. Gomez, M.L. Wolf, B.M. Tibbetts, and T.H. March. 2011b. The acute toxicity, tissue distribution,
and histopathology of inhaled ricin in Sprague Dawley rats and BALB/c mice. Inhal. Toxicol. 23(5):247-256.
24
Prepublication Copy

Bch, T.R., E.A. Schfer, M.T. Demmel, I. Boekhoff, H. Thiermann, T. Gudermann, D. Steinritz, and A. Schmidt.
2013. Functional expression of the transient receptor potential channel TRPA1, a sensor for toxic lung inhalants, in pulmonary epithelial cells. Chem. Biol. Interact. 206(3):462-471.
Chen, Y., C. Guo, L. Lim, S. Cheong, Q. Zhang, K. Tang, and J. Reboud. 2008. Compact microelectrode array system: Tool for in situ monitoring of drug effects on neurotransmitter release from neural cells. Anal Chem.
80(4):1133-1140.
Cui, H.F., J.S. Ye, Y. Chen, S.C. Chong, and F.S. Sheu. 2006. Microelectrode array biochip: Tool for in vitro drug
screening based on the detection of a drug effect on dopamine release from PC12 cells. Anal Chem.
78(18):6347-6355.
deLemos, A.S., D.M. Foureau, C. Jacobs, W. Ahrens, M.W. Russo, and H.L. Bonkovsky. 2014. Drug-induced liver
injury with autoimmune features. Semin. Liver Dis. 34(2):194-204.
DHHS (US Department of Health & Human Services). 2014. Chemical Warfare. Classes of Chemical Agents.
Specialized Information System [online]. Available: http://sis.nlm.nih.gov/enviro/chemicalwarfare.html#a1
[accessed March 13, 2015].
Dunlop, J., R. Roncarati, B. Jow, H. Bothmann, T. Lock, D. Kowal, M. Bowlby, and G.C. Terstappen. 2007. In vitro
screening strategies for nicotinic receptor ligands. Biochem. Pharmacol. 74(8):1172-1181.
Falk, M., M. Hausmann, E. Lukov, A. Biswas, G. Hildenbrand, M. Davdkov, E. Krasavin, Z. Kleibl, I. Falkov, L. Jekov, L. tefankov, J. evk, M. Hofer, A. Bakov, P. Matula, A. Boreyko, J. Vachelov, A.
Michaelidesov, and S. Kozubek. 2014. Determining Omics spatiotemporal dimensions using exciting new
nanoscopy techniques to assess complex cell responses to DNA damage: Part A--radiomics. Crit. Rev. Eukaryot. Gene Expr. 24(3):205-223.
Gutzkow, K.B., T.M. Langleite, S. Meier, A. Graupner, A.R. Collins, and G. Brunborg. 2013. High-throughput
comet assay using 96 minigels. Mutagenesis 28(3):333-340.
Hill, A.J., N.A. Jones, I. Smith, C.L. Hill, C.M. Williams, G.J. Stephens, and B.J. Whalley. 2014. Voltage-gated
sodium (NaV) channel blockade by plant cannabinoids does not confer anticonvulsant effects per se. Neurosci.
Lett. 566:269-274.
Himmel, H.M. 2013. Drug-induced functional cardiotoxicity screening in stem cell-derived human and mouse cardiomyocytes: Effects of reference compounds. J. Pharmacol. Toxicol. Methods 68(1):97-111.
Holstege, C.P., L.K. Bechtel, T.H. Reilly, B.P. Wispelwey, and S.G. Dobmeier. 2007. Unusual but potential agents
of terrorists. Emerg. Med. Clin. North Am. 25(2):549-566.
Huh, D., D.C. Leslie, B.D. Matthews, J.P. Fraser, S. Jurek, G.A. Hamilton, K.S. Thorneloe, M.A. McAlexander, and
D.E. Ingber. 2012. A human disease model of drug toxicity-induced pulmonary edema in a lung-on-a-chip microdevice. Sci. Transl. Med. 4(159):159ra147.
Jellett, J.F., L.J. Marks, J.E. Stewart, M.L. Dorey, W. Watson-Wright, and J.F. Lawrence. 1992. Paralytic shellfish
poison (saxitoxin family) bioassays: Automated endpoint determination and standardization of the in vitro tissue culture bioassay, and comparison with the standard mouse bioassay. Toxicon. 30(10):1143-1156.
Jensen, K.H., and C. Rekling. 2010. Development of a no-wash assay for mitochondrial membrane potential using
the styryl dye DASPEI. J. Biomol. Screen. 15(9):1071-1081.
Jin, S., K.S. Sarkar, Y.N. Jin, Y. Liu, D. Kokel, T.J. Van Ham, L.D. Roberts, R.E. Gerszten, C.A. Macrae, and R.T.
Peterson. 2013. An in vivo zebrafish screen identifies organophosphate antidotes with diverse mechanisms of
action. J. Biomol. Screen. 18(1):108-115.
Jones, J.D., H.L. Kirsch, R.L. Wortmann, and M.H. Pillinger. 2014. The causes of drug-induced muscle toxicity.
Curr. Opin. Rheumatol. 26(6):697-703.
Kelesidis, T., C.K. Roberts, D. Huynh, O. Martnez-Maza, J.S. Currier, S.T. Reddy, and O.O. Yang. 2014. A high
throughput biochemical fluorometric method for measuring lipid peroxidation in HDL PLoS One
9(11):e111716.
Li, Y., X. Feng, W. Du, Y. Li, and B.F. Liu. 2013. Ultrahigh-throughput approach for analyzing single-cell genomic
damage with an agarose-based microfluidic comet array. Anal. Chem. 85(8):4066-4073.
Liu, S., Y. Zhang, M. Moayeri, J. Liu, D. Crown, R.J. Fattah, A.N. Wein, Z.X. Yu, T. Finkel, and S.H. Leppla.
2013. Key tissue targets responsible for anthrax-toxin-induced lethality. Nature 501(7465):63-68.
Lopez-Izquierdo, A., M. Warren, M. Riedel, S. Cho, S. Lai, R.L. Lux, K.W. Spitzer, I.J. Benjamin, M. TristaniFirouzi, and C.J. Jou. 2014. A near-infrared fluorescent voltage-sensitive dye allows for moderate-throughput
electrophysiological analyses of human induced pluripotent stem cell-derived cardiomyocytes. Am. J. Physiol.
Heart Circ. Physiol. 307(9):H1370-H1377.
Prepublication Copy
25
Mioulane, M., G. Foldes, N.N. Ali, M.D. Schneider, and S.E. Harding. 2012. Development of high content imaging
methods for cell death detection in human pluripotent stem cell-derived cardiomyocytes. J. Cardiovasc.
Transl. Res. 5(5):593-604.
Monteiro-Riviere, N.A., and A.O. Inman. 1997. Ultrastructural characterization of sulfur mustard-induced vesication
in isolated perfused porcine skin. Microsc. Res. Tech. 37(3):229-241.
Morrison, K.C., and P.J. Hergenrother. 2012. Whole cell microtubule analysis by flow cytometry. Anal. Biochem. 420(1):26-32.
NRC (National Research Council). 2005. Impact of Revised Airborne Exposure Limits on Non-Stockpile Chemical
Materiel Program Activities. Washington, DC: National Academies Press.
NRC (National Research Council). 2007. Toxicity Testing in the 21st Century: A Vision and a Strategy. Washington, DC: National Academies Press.
Pantazides, B.G., B.S. Crow, J.W. Garton, J.A. Quiones-Gonzlez, T.A. Blake, J.D. Thomas, and R.C. Johnson.
2015. Simplified method for quantifying sulfur mustard adducts to blood proteins by Ultrahigh Pressure Liquid Chromatography-Isotope Dilution Tandem Mass Spectrometry. Chem. Res. Toxicol. 28(2):256-261.
Pieperhoff, S., K.S. Wilson, J. Baily, K. de Mora, S. Maqsood, S. Vass, J. Taylor, J. Del-Pozo, C.A. MacRae, J.J.
Mullins, and M.A. Denvir. 2014. Heart on a plate: Histological and functional assessment of isolated adult
zebrafish hearts maintained in culture. PLoS One 9(5):e96771.
Pointon, A., N. Abi-Gerges, M.J. Cross, and J.E. Sidaway. 2013. Phenotypic profiling of structural cardiotoxins in
vitro reveals dependency on multiple mechanisms of toxicity. Toxicol. Sci. 132(2):317-326.
Pointon, A., A.R. Harmer, I.L. Dale, N. Abi-Gerges, J. Bowes, C. Pollard, and H. Garside. 2015. Assessment of cardiomyocyte contraction in human-induced pluripotent stem cell-derived cardiomyocytes. Toxicol Sci.
144(2):227-237.
Prasad, R.Y., J.K. McGee, M.G. Killius, D.A. Suarez, C.F. Blackman, D.M. DeMarini, and S.O. Simmons. 2013.
Investigating oxidative stress and inflammatory responses elicited by silver nanoparticles using highthroughput reporter genes in HepG2 cells: effect of size, surface coating, and intracellular uptake. Toxicol. In
Vitro 27(6):2013-2021.
Riviere, J.E., J.D. Brooks, P.L. Williams, and N.A. Monteiro-Riviere. 1995. Toxicokinetics of topical sulfur mustard
penetration, disposition, and vascular toxicity in isolated perfused porcine skin. Toxicol. Appl. Pharmacol.
135(1):25-34.
Sakamuru, S., X. Li, M.S. Attene-Ramos, R. Huang, J. Lu, L. Shou, M. Shen, R.R. Tice, C.P. Austin, and M. Xia.
2012. Application of a homogenous membrane potential assay to assess mitochondrial function. Physiol. Genomics 44(9):495-503.
Scott, C.W., X. Zhang, N. Abi-Gerges, S.D. Lamore, Y.A. Abassi, and M.F. Peters. 2014. An impedance-based cellular assay using human iPSC-derived cardiomyocytes to quantify modulators of cardiac contractility. Toxicol
Sci. 142(2):331-338.
Sharma, R., B. Gupta, N. Singh, J.R. Acharya, K. Musilek, K. Kuca, and K.K. Ghosh. In press. Development and
structural modifications of cholinesterase reactivators against chemical warfare agents in last decade: A review. Mini Rev. Med. Chem.
Sirenko, O., E.F. Cromwell, C. Crittenden, J.A. Wignall, F.A. Wright, and I. Rusyn. 2013. Assessment of beating
parameters in human induced pluripotent stem cells enables quantitative in vitro screening for cardiotoxicity.
Toxicol. Appl. Pharmacol. 273(3):500-507.
Sirenko, O., J. Hesley, I. Rusyn, and E.F. Cromwell. 2014a. High-content assays for hepatotoxicity using induced
pluripotent stem cell-derived cells. Assay Drug Dev. Technol. 12(1):43-54.
Sirenko, O., J. Hesley, I. Rusyn, and E.F. Cromwell. 2014b. High-content high-throughput assays for characterizing
the viability and morphology of human iPSC-derived neuronal cultures. Assay Drug Dev. Technol. 12(910):536-547.
Steinhoff, R.F., M. Ivarsson, T. Habicher, T.K. Villiger, J. Boertz, J. Krismer, S.R. Fagerer, M. Soos, M. Morbidelli,
M. Pabst, and R. Zenobi. 2015. High-throughput nucleoside phosphate monitoring in mammalian cell fedbatch cultivation using quantitative matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Biotechnol J. 10(1):190-198.
Stenger, B., F. Zehfu, H. Mckter, A. Schmidt, F. Balszuweit, E. Schfer, T. Bch, T. Gudermann, H. Thiermann,
and D. Steinritz. In press. Activation of the chemosensing transient receptor potential channel A1 (TRPA1) by
alkylating agents. Arch. Toxicol.
Sturm, M.B., and V.L. Schramm. 2009. Detecting ricin: Sensitive luminescent assay for ricin A-chain ribosome
depurination kinetics. Anal. Chem. 81(8):2847-2853.
26
Prepublication Copy

Suzuki, O.T., A. Frick, B.B. Parks, O.J. Trask Jr., N. Butz, B. Steffy, E. Chan, D.K. Scoville, E. Healy, C. Benton,
P.E. McQuaid, R.S. Thomas, and T. Wiltshire. 2014. A cellular genetics approach identifies gene-drug interactions and pinpoints drug toxicity pathway nodes. Front. Genet. 5:272.
Tenberken, O., J. Mikler, I. Hill, K. Weatherby, H. Thiermann, F. Worek, and G. Reiter. 2010. Toxicokinetics of
tabun enantiomers in anaesthetized swine after intravenous tabun administration. Toxicol. Lett. 198(2):177181.
Thoren, T.M., K.S. Thompson, P.S. Cardona, A.K. Chaturvedi, and D.V. Canfield. 2013. In vitro absorption of atmospheric carbon monoxide and hydrogen cyanide in undisturbed pooled blood. J. Anal. Toxicol. 37(4):203207.
Thorn, J. 2001. The inflammatory response in humans after inhalation of bacterial endotoxin: A review. Inflamm.
Res. 50(5):254-261.
Vale, C., A. Alfonso, M.R. Vieytes, X.M. Romars, F. Arvalo, A.M. Botana, and L.M. Botana. 2008. In vitro and
in vivo evaluation of paralytic shellfish poisoning toxin potency and the influence of the pH of extraction.
Anal. Chem. 80(5):1770-1776.
Vallet, V., C. Cruz, J. Licausi, A. Bazire, G. Lallement, and I. Boudry. 2008. Percutaneous penetration and distribution of VX using in vitro pig or human excised skin validation of demeton-S-methyl as adequate simulant for
VX skin permeation investigations. Toxicology 246(1):73-82.
van der Linden, S.C., A.R. von Bergh, B.M. van Vught-Lussenburg, L.R. Jonker, M. Teunis, C.A. Krul, and B. van
der Burg. 2014. Development of a panel of high-throughput reporter-gene assays to detect genotoxicity and
oxidative stress. Mutat. Res. Genet. Toxicol. Environ. Mutagen. 760:23-32.
Vangveravong, S., E. McElveen, M. Taylor, J. Xu, Z. Tu, R.R. Luedtke, and R.H. Mach. 2006. Synthesis and characterization of selective dopamine D2 receptor antagonists. Bioorg. Med. Chem. 14(3):815- 825.
Vongs, A., K.J. Solly, L. Kiss, D.J. Macneil, and C.I. Rosenblum. 2011. A miniaturized homogenous assay of mitochondrial membrane potential. Assay Drug Dev. Technol. 9(4):373-381.
Wasalathanthri, D.P., S. Malla, I. Bist, C.K. Tang, R.C. Faria, and J.F. Rusling. 2013. High-throughput metabolic
genotoxicity screening with a fluidic microwell chip and electrochemiluminescence. Lab. Chip. 13(23):45544562.
Watson, C., J. Ge, J. Cohen, G. Pyrgiotakis, B.P. Engelward, and P. Demokritou. 2014. High-throughput screening
platform for engineered nanoparticle-mediated genotoxicity using CometChip technology. ACS Nano. 8(3):
2118-2133.
Wijte, D., M.J. Alblas, D. Noort, J.P. Langenberg, and H.P. van Helden. 2011. Toxic effects following phosgene exposure of human epithelial lung cells in vitro using a CULTEX system. Toxicol. In Vitro 25(8):2080-2087.
Wille, T., H. Thiermann, and F. Worek. 2010. Development of a high-throughput screening for nerve agent detoxifying materials using a fully-automated robot-assisted biological assay. Toxicol. In Vitro 24(3):1026-1031.
Wills, L.P., G.C. Beeson, R.E. Trager, C.C. Lindsey, C.C. Beeson, Y.K. Peterson, and R.G. Schnellmann. 2013.
High-throughput respirometric assay identifies predictive toxicophore of mitochondrial injury. Toxicol. Appl.
Pharmacol. 272(2):490-502.
Worek, F., P. Eyer, N. Aurbek, L. Szinicz, and H. Thiermann. 2007. Recent advances in evaluation of oxime efficacy in nerve agent poisoning by in vitro analysis. Toxicol. Appl. Pharmacol. 219(2-3):226-234.
Xiao, J., R.B. Free, E. Barnaeva, J.L. Conroy, T. Doyle, B. Miller, M. Bryant-Genevier, M.K. Taylor, X. Hu, A.E.
Dulcey, N. Southall, M. Ferrer, S. Titus, W. Zheng, D.R. Sibley, and J.J. Marugan. 2014. Discovery, optimization, and characterization of novel D2 dopamine receptor selective antagonists. J. Med. Chem. 57(8):34503463.
Yakushenko, A., E. Ktelhn, and B. Wolfrum. 2013. Parallel on-chip analysis of single vesicle neurotransmitter
release. Anal. Chem. 85(11):5483-5490.
Zielonka, J., G. Cheng, M. Zielonka, T. Ganesh, A. Sun, J. Joseph, R. Michalski, W.J. O'Brien, J.D. Lambeth, and
B. Kalyanaraman. 2014. High-throughput assays for superoxide and hydrogen peroxide: Design of a screening
workflow to identify inhibitors of NADPH oxidases. J. Biol. Chem. 289(23):16176-16189.
Prepublication Copy
27
3
Nontesting Approaches Relevant to
Prediction of Acute Toxicity and Potency
The term nontesting approaches was coined during the development of the European Union Registration, Evaluation, Authorization, and Restriction of Chemicals (REACH) regulation (EU 2006; ECHA
2008) to include the search and retrieval of existing data, the identification of structural alerts1 to indicate
activity, the grouping of chemicals for read-across, and the development and application of quantitative
structure-activity models. In practice, nontesting approaches are used to accomplish various tasks. For
example, predictions based on structureactivity relationships (SARs) and quantitative structureactivity
relationships (QSARs) are used to fill specific data gaps in lieu of experimental testing, to support findings or conclusions in integrated chemical assessments, and to substantiate predictions of various properties for structurally related chemicals. A key assumption that underpins nontesting approaches is that the
property of a chemical with respect to how it will interact with a defined biological system is inherent in
its molecular structure; thus, similar chemicals should have similar biological activities (the similarity
principle) (Raunio 2011). More detailed information about nontesting approaches can be found in Cronin
and Madden (2010).
This chapter discusses nontesting approaches in the context of the conceptual framework described
in Chapter 2. Box 3-1 provides definitions for some of the terms used in the chapter. The chapter discusses the use of available data to characterize chemicals of interest and in silico approaches to predict physical hazards, chemical reactivity, pharmacokinetic properties, and acute toxicity. It also provides selected
examples to demonstrate how computational tools could be used in predictive toxicology.
INITIAL CHEMICAL CHARACTERIZATION
The starting point in the application of any nontesting approach for predicting acute toxicity is a preliminary search and evaluation of available data on the chemical of interest. The effort often begins with
database queries, literature searches, and other approaches for finding information about the chemicals
structure, physicochemical properties, and acute toxicity (see Box 3-2). The committee notes that the
forthcoming REACH regulation requires in vivo acute oral toxicity information for chemicals manufactured or imported into Europe at greater than 1 metric ton per year (ECHA 2012). It is anticipated that a
large volume of in vivo oral data will be potentially disseminated publically after the REACH May 2018
deadline.
The available information can help in identifying a chemicals potential for direct physical hazards,
most relevant routes of exposure, likely bioavailability, and potential for inducing (human) toxicity (NRC
2014). It should also be considered before designing or initiating new in vitro or in vivo experimental
studies, in interpreting existing empirical data, or in selecting appropriate (Q)SAR models.2
1
2
A structural alert is a chemical structure that has been linked to toxicity or a specific toxicity end point.
The committee uses the shorthand notation (Q)SAR to indicate both SAR and QSAR.
28
Prepublication Copy
Nontesting Approaches Relevant to Prediction of Acute Toxicity and Potency
BOX 3-1 Definitions of Selected Nontesting Approaches

A structureactivity relationship (SAR) is a qualitative association between a chemical (sub)structure
(such as a functional group) and the potential of a chemical that contains the (sub)structure to exhibit a
particular biological effect.
A quantitative structureactivity relationship (QSAR) is a mathematical relationship between a quantifiable aspect of chemical structure and a chemical property or reactivity or a well defined biological
activity, such as toxicity (EPA 2012). QSARs can be derived to predict quantitative or qualitative end
points.
A quantitative structureproperty relationship (QSPR) is a special case of QSAR in which a physicochemical property is modelled as the response variable.
An expert system is a software tool that specifically encodes compilations of SARs, QSARs, or both to
enable rational predictions of toxicity to be made on the basis of structure alone. Expert systems are
typically categorized as statistical (for example, TOPKAT and Accelrys Inc), knowledge-based (for example, such SAR-based approaches as Derek Nexus and LHASA Ltd), or hybrid (for example,
TIMES-SS).
Category approach, analogue approach, and read-across: Category and analogue approaches are
techniques for grouping chemicals; read-across is a technique for filling data gaps in category and analogue approaches (ECHA 2008; OECD 2014). Read-across can be qualitative or quantitative and
uses existing information on the property of a substance (source chemical)to make a prediction of
the same property for another substance (target chemical) that is considered similar with respect to
the end point of interest (Worth 2008). Analogue approaches are used for grouping a small number of
chemicals when there are no apparent trends in properties.
A chemical category is a group of chemicals whose physicochemical and human healthor environmental toxicological propertiesor environmental-fate properties are likely to be similar or follow a
regular pattern as a result of structural similarity (OECD 2007a, 2014). Chemical similarity could be
based on a variety of properties, including the presence of a common functional group (such as an
aldehyde), common constituents or chemical classes, similar carbon range numbers, or common precursors or breakdown products.
In silico approaches include computational modeling, SAR analysis, analysis of physicochemical characteristics, and read-across techniques.
As described below, physicochemical properties of interest in predictive toxicology can be nominally categorized into three broad types: physical properties, solvation properties, and molecular attributes
(NRC 2014). There are methods for empirically measuring the properties and in silico approaches for estimating their values (see Box 3-3). It can be particularly helpful to complement estimated values with
experimental measurements, when that is possible.
Physical properties. Physical properties include such characteristics as freezing point, melting
point, boiling point, vapor pressure, and viscosity. Melting point, boiling point, and vapor pressure can be
used to predict a chemicals likely physical state, which is pertinent in determining the most relevant route
of exposure for any testing or indeed what practical challenges might need to be overcome in in vitro testing
scenarios or even what issues to consider in interpreting in vivo results and associated testing protocols.
Solvation properties. Solvation properties describe a chemicals interaction with different phases
and its interaction between phases (for example, logKow represents the partitioning between octanol and
water). Water solubility and logKow are particularly helpful in determining the technical feasibility of performing in vitro test protocols given that they typically use aqueous media.
Prepublication Copy
29
Molecular attributes. Molecular attributes capture properties related to molecular shape and size.
Electronic characteristics of molecules, such as frontier orbital energies and polarizability, that are related
to reactivity could be considered to constitute a type of molecular attribute. They also play a role in predicting likely bioavailability and toxicity.
BOX 3-2 Primary Data Considered During a Preliminary Characterization of a Chemical of Interest
Chemical structure: Chemical structure is the spatial arrangement of a molecules constituent atoms.
PubChem, DSSTox, and ChemIDplus are examples of searchable chemical-structure databases.
Physicochemical properties: Physicochemical properties contribute to the inherent hazards posed by a
chemical, including its ability to interfere with normal biological processes. Physicochemical properties
could also define a chemical's physical hazards of interest (such as corrosivity). Physical properties
include freezing point, boiling point, melting point, infrared spectrum, electronic characteristics, viscosity, and density. Other properties of relevance include solvation properties, such as phase partitioning
and solubility. One of the more important phase-partition coefficients is obtained from a system in
which one solvent is water or an aqueous phase and the second is organic and hydrophobic, such as
1-octanol, that is, the octanolwater partition coefficient (Kow). An example of a database source of
physicochemical-property data is the National Institute of Standards and Technology Chemistry WebBook; another is the PHYSPROP database, which is integrated into the US Environmental Protection
Agencys EPISuite software.
Acute toxicity (for example, rodent LD50 or LC50 values): These data might be available from primary
sources (such as peer-reviewed literature) and secondary sources (such as the Merck Handbook). By
far the most convenient sources of data are compiled databases that are readily searchable by chemical identifiers, such as chemical name, Chemical Abstracts Service registry number, or chemical structure. An example is the National Library of Medicine TOXNET database.
See Appendix B for additional information and examples.
BOX 3-3 In Silico Approaches for Predicting Physicochemical Properties

In the absence of data on physicochemical properties, reasonable estimates based on chemical
structure are feasible with the use of QSAR and QSPR models (Dearden and Worth 2007). A discussion of those methods is beyond the scope of the present activity. However, the National Research
Council report A Framework to Guide Selection of Chemical Alternatives provides a succinct discussion of published models that can be used to characterize a number of physicochemical properties
(NRC 2014). That report discusses methods used to estimate molecular hydrophobicity (or lipophilicity) and other physicochemical end points, such as aqueous solubility, pKa, and the electronic properties of molecules.
A number of software packages and algorithms are available for predicting physicochemical properties, and predictions are often in excellent agreement with experimentally derived values. For example, pKa can be estimated by using Taft and Hansch fragment coefficients, and Perrin et al. (1981)
contains an extensive compilation of the fragment values and relevant equations to do so. For convenience, software tools, such as SPARC or those created by ACD Labs, contain algorithms for estimating pKa directly from chemical structure. The user of such tools, however, must have a basic understanding of the inherent advantages and limitations of the various algorithms as they are related to the
accuracy of physicochemical-property prediction. In general, the QSAR models available for prediction
of the key physicochemical characteristics are best suited for small organic chemicals that typically
have one functional group. Other chemicalssuch as pesticides, drug-like chemicals, or other pharmacological activestypically lack published data to derive such QSARs.
30
Prepublication Copy

USE OF PHYSICOCHEMICAL PROPERTIES TO PREDICT PHYSICAL
HAZARDS, CHEMICAL REACTIVITY, AND PHARMACOKINETICS
Physicochemical data can be used to predict a chemicals physical hazard, reactivity, and pharmacokinetics, including absorption by different exposure routes, distribution in the body, and likely metabolites.
Approaches that apply knowledge about a chemicals physicochemical properties to predictive toxicology
presume that for a chemical to exert a toxic effect, it typically must be bioavailable to such an extent that it
(or its metabolite) reaches a biochemical target, where it can exert its toxic effect (Meek et al. 2013).
Use of Physicochemical Properties to Predict Physical
Hazard and Chemical Stability or Reactivity
Assessing the likely irritant or corrosive effects of a chemical would be a helpful component of a tiered
evaluation strategy for predicting acute toxicity.3 In the absence of measured irritant or corrosive data, a
handful of (Q)SAR approaches are useful in identifying potential irritants or corrosives. The German Federal Institute for Risk Assessment (BfR) rule base (Gerner et al. 2004; Hulzebos et al. 2005) is one example of
an expert system that uses physicochemical exclusion rules and structural alert inclusion rules to determine
likely skin or eye irritation hazard.4 The BfR rule base has been encoded into software tools, including the
Organisation for Economic Co-operation and Development (OECD) QSAR Toolbox5 and the European
Commission Joint Research Centre Toxtree (Rorije and Hulzebos 2005; Tsakovska et al. 2007). Some
QSARs published for specific chemical classes have relied on such properties as pKa, dipole moment,
logKow, and molecular weight or volume to estimate likely irritation potential and potency (Barratt 1996).
Both pKa and dipole moment have been found to be useful measures for modeling chemical reactivity depending on whether the target substance is an acid, a base, or a neutral organic; and logKow and molecular
weight have served as useful surrogates for modeling partitioning. QSARs also exist within expert systems,
such as TOPKAT and MCASE, for the prediction of irritation or corrosion. Saliner et al. (2008) reviewed
the status of (Q)SAR approaches for irritation and corrosion.
Consideration should also be paid to the inherent stability or electrophilic reactivity of the chemical. A
chemical might exert its effects in its parent form or be transformed abiotically or biotically to a metabolite
that is a more relevant target for evaluation. Some substances are rapidly hydrolyzed; for example, acid
chlorides and acid anhydrides are rapidly hydrolyzed to their corresponding carboxylic acids. Other substances are capable of being oxidized when exposed to air; for example, p-hydroquinone is rapidly oxidized
to its corresponding benzoquinone, which is highly reactive. The OECD Toolbox contains simulators that
help in predicting such transformations. Consideration of how a chemical might be transformed is important
in interpreting experimental data, performing new testing, or using the most relevant target for (Q)SAR
analyses.
Skin irritation or corrosion can be investigated in vitro by virtue of assays, such as Corrositex (OECD 2006) for
corrosion and EpiDerm (OECD 2013a) for irritation. For eye irritation or corrosion, various ex vivo and in vitro
assays are available, including the bovine corneal opacity permeability test (OECD 2013b), the isolated chicken-eye
test (OECD 2013c), or the EpiOcular eye-irritation test method. A tabulation of assays for irritation and corrosion
that have been validated by ECVAM or ICCVAM or that exist as test guidelines under OECD are provided on the
AltTox.org Web site (AltTox 2014) and are discussed in Chapter 4.
4
The BfR rule base combines two approaches: exclusion rules that use physicochemical thresholds to identify
chemicals that are not skin irritants or corrosive and inclusion rules that use structural alerts to identify chemicals
that are potentially irritants or corrosive (Saliner et al. 2007). The rule base assigns a regulatory classification for
skin or eye irritation or corrosion.
5
The committee refers here to the toolbox by its official name rather than OECD (Q)SAR Toolbox, which would
be more appropriate because the Toolbox includes both SAR and QSAR approaches.
Prepublication Copy
31
Use of Physicochemical Properties to Predict Chemical Disposition and Metabolism
Pharmacokinetics describes the disposition of a chemical in an organism and considers chemical absorption, distribution, metabolism, and excretion (ADME). Pharmacokinetic properties can play an important role in the assessment of a chemicals effects on or risks to the body. In silico approaches have been
developed to predict many ADME processes; the sections below focus largely on approaches that are directly relevant for predicting acute toxicity.
Absorption: Oral
Some physicochemical propertiessuch as molecular weight, the number of hydrogen-bond donors
and acceptors, and logKowhave been shown to be predictive of oral absorption. For example, Lipinskis
rule of 5 is considered helpful in evaluating the likely absorption, permeability, and toxicity of drug-like
substances (Lipinski et al. 2001) and considers the three properties noted to make predictions about chemical behavior. Other examples of heuristic rules are provided in Table 3-1.
In addition to heuristic rules, several QSAR models have been developed to determine intestinal absorption and oral absorption. Iyer et al. (2007) used a membrane-interaction QSAR analysis to build models for human oral intestinal drug absorption. Castillo-Garit et al. (2008) developed a mathematical model
that used linear indexes to predict the in vitro permeability of 157 chemicals in a Caco-2 cell model. Their
mathematical model had greater than 80% accuracy in predicting how well a drug would be absorbed by
Caco-2 cells. Guerra et al. (2010) developed an artificial neural network by using CODES 2D descriptions to predict oral drug absorption. Suenderhauf et al. (2011) used a broad selection of machine learning
and statistical methods to derive classification and prediction models for human intestinal absorption.
Several recent reviews discuss the status of such QSAR models (Xu and Mager 2011; Silva and Trossini
2014; Wang and Hou in press).
There are also physiologically based packages that can predict oral absorption rates of drugs, such as
GastroPlus and SimCyp (Kuentz et al. 2006; Rostami-hodjegan and Tucker 2007; Yang et al. 2007; Simulations Plus 2010; Grbic et al. 2011). However, if the goal is to determine likely oral absorption to help to
prioritize chemicals, the heuristic rules described above might be adequate for that task.
Absorption: Dermal
LogKow, water solubility, and molecular weight are also useful inputs for estimating dermal absorption characteristics of a target substance. QSAR models for predicting the dermal permeability coefficient
(Kp)a measure useful for modeling dermal penetrationtypically rely on LogKow and molecular weight
as input variables. Potts and Guy (1992) derived such a model that is also encoded in DERMWIN as part
of EPAs EpiSuite software. Over the years, the model derived by Potts and Guy has been modified to
address limitations, and many variants now exist. Mitragotri et al. (2011) reviewed the status of models
for the prediction of skin permeability in terms of their strengths, limitations, and future prospects.
TABLE 3-1 Examples of Heuristic Rules to Predict Oral Absorption
Rule
GlaxoSmithKline
rule of 4/400
Descriptiona
Chemicals with cLogP < 4 and MW < 400 Da have superior drug-like
properties compared with chemicals with cLogP > 4 and MW > 400 Da
Reference
Gleeson 2008
Pfizer rule of 3/75
Chemicals with cLogP > 3 and total PSA < 75 are 2.5 times more likely
to have in vivo toxicity than ones with cLogP < 3 and total PSA > 75
Hughes et al. 2008
AstraZeneca
Alkalinity and increased cLogP are associated with multiple

positive responses in various toxicity assays
Leeson and Springthorpe 2007
cLogP is the name of a software program that generates an estimate of logKow.

Abbreviations: MW, molecular weight; PSA, polar surface area.
32
Prepublication Copy

Other researchers have incorporated additional information, such as degree of hydrogen bonding and
melting point, to refine skin penetration estimates (Hostnek 1997; Magnusson et al. 2004; ten Berge 2009;
Dancik et al. 2013a). Models by ten Berge (2009) and Dancik et al. (2013a,b) are helpful in evaluating systemic availability as a result of dermal exposure and thus provide a means of extrapolating a dermal acutetoxicity (LD50) value from an oral acute-toxicity (LD50) value. Kastings model (as discussed in Dancik et al.
2013a,b) explicitly takes into account various components of the skin structure, including the stratum
corneum, viable epidermis, and dermis. The model simulates one-dimensional transient passive transport
into a skin slab. Some properties are also required as inputs for a simulation, including logKow, vapor pressure, melting and boiling points, molecular weight, chemical class (alcohol, hydrocarbon, or other organic),
and presence of a pharmacophore6 as defined by Yamazaki and Kanaoka (2004). The model is available for
use from the National Institute for Occupational Safety and Health Web site (NIOSH 2013).
Absorption and Deposition: Inhalation
For nonvolatile chemicals, particle size is an important consideration because it affects deposition in
the respiratory tract and influences whether a particle poses an inhalation hazard (ECETOC 2012; Brown
et al. 2013). Brown et al. (2013) predicted that about half of all 10-m particles penetrate into the thorax
and that about 20% or less of all 10-m particles would penetrate to the extrathoracic airways and into the
lower respiratory tract.
For volatile substances, physicochemical characteristicssuch as vapor pressure, water solubility,
and reactivityare also important for predicting acute toxicity by the inhalation route (Veith and Wallace
2006; Veith et al. 2009).
Metabolism
Several tissuesincluding the lung, skin, liver, intestine, and kidneyhave enzymes that can convert a parent chemical to a metabolite, for example, through oxidation and conjugation processes. Whereas parent chemical metabolism typically results in a more hydrophilic chemical that is more easily excreted, a reactive toxic metabolite is sometimes formed. Not considering that possibility and focusing solely
on the parent chemical will therefore be inadequate in characterizing a chemicals potential to elicit acute
toxicity accurately.
Predicting chemical metabolism requires tools that can identify the functional groups or structural
components of the parent chemical that are vulnerable to metabolism (sites of metabolism) and the structure of possible metabolites. Additional information about enzyme structure and function and about the
effect of metabolism on the induction or inhibition of metabolizing enzymes could also be considered, but
for the purposes of predicting the potential of chemicals to elicit acute toxicity, this discussion will focus
primarily on the first two factors. Determining the site of metabolism allows prediction of overall metabolic stability, such as the rate of activity (Vmax, Km) or clearance rate (Clint), that is measured either in
vivo or in vitro. Predicting metabolic structures involves listing possible metabolites from reactions that
the chemical could undergo (biotransformation).
Metabolism-predictive tools are based on large compilations of databases derived from metabolism
information in the literature, for example, Accelrys Metabolite Database, Metabolite, MetaBase, and
MetaDrug (Kirchmair et al. 2012). The metabolism information in a database can be used to identify likely metabolic sites on the basis of what is known about the target chemical structure or a similar chemical
structure. The databases can also be used to predict possible metabolites by using information that describes enzyme activity, such as binding pocket sites. Most available metabolism-predictive tools consider
6
A pharmacophore is the collection of steric and electrostatic features of different chemicals that are necessary to
ensure optimal molecular interactions with a specific biological target (Langer and Wolber 2004).
Prepublication Copy
33
a single aspect of metabolic reactionssuch as the reaction energy barrier, geometrical properties, or
pharmacokinetic propertiesto predict sites of metabolism or potential metabolites. ADMET Predictor is
an example of a commercial product that uses the Accelrys Metabolite Database to predict various metabolic stability values for a series of cytochromes (Simulations Plus 2010). Kirchmair et al. (2012) provide
a comprehensive overview of methods for predicting sites of metabolism.
Knowledge-driven approaches, such as expert systems, allow extrapolation of structure to likely metabolites by using advanced reasoning rules and expert-system metabolite ranking, such as MetabolExpert, META, Meteor, and Metaprint2D-React. However, one problem is that they can generate a large
number of metabolites from which it is difficult to determine which metabolites are the relevant and stable ones that should be considered. Other metabolism-predictive tools introduce the expert-system features with more refined computational algorithms to support the decision method and therefore limit the
number of metabolites that are generated. Indeed, the software program Tissue Metabolism Simulator
(TIMES) uses a comprehensive library of biotransformation information and a heuristic algorithm to generate plausible metabolic maps that are relevant to specific end points, such as skin sensitization or genotoxicity (Mekenyan et al. 2012; Patlewicz et al. 2014). Many of the TIMES metabolism simulators have
been made freely available in the OECD QSAR Toolbox.
Limitations and Need for Improvement
Many tools for predicting physicochemical properties that are relevant for the evaluation of chemical disposition and distribution factors are available, but they are limited by their training sets.7 Such tools
are generally most applicable for small organic chemicalschemicals that have molecular weights of 500
Da or less (that is, not mixtures or polymers).
In vitro and in silico predictions of absorption for various routes of exposure are still crude, and current models might have little applicability to the Department of Defense (DOD).8 For oral absorption,
Lipinskis rule of 5, which is based on experience in the drug-discovery world, might provide a convenient set of heuristics for chemicals of interest to DOD but would need to be evaluated to determine its applicability. The prediction of dermal permeability at the simplest level is illustrated by QSARs that predict
logKp or logJmax with such inputs as molecular weight and logKow as exemplified by Potts and Guy
(1992) or Magnusson et al. (2004) (see also Fitzpatrick et al. 2004 and Mitragotri et al. 2011). Although
refinements have been made to simulate penetration or systemic bioavailability (Dancik et al. 2013a,b),
the underlying characteristics are still based largely on the heterogeneous dataset first compiled by Flynn
(1990), which is limited in its coverage of chemicals.
Current metabolism-predictive approaches have several limitations. First, most of the existing metabolism-predictive tools were designed primarily to inform drug development. There are few examples in
which such modeling tools have been used to evaluate volatile or lipophilic chemicals (Peyret and Krishnan 2012; Kirman et al. in press). Thus, the available metabolism training sets will need to be expanded
for the chemicals of interest. Second, although metabolism-predictive tools adequately predict transformations of various chemicals, they do a poor job of distinguishing differences in reactivity of closely related structural analogues. In most cases, the tools can only estimate the reactivity of the individual molecular sites. As a result, they have limited use for prioritizing a broad set of structurally related chemicals
(Kirchmair et al. 2012). One interim solution that DOD might consider is to evaluate metabolites with
7
Training sets are data that are used to develop predictive models or tools.
For example, in vitro assays for absorption (such as Caco-2 monolayer crossing) were developed primarily to
predict systemic absorption after deliberate oral dosing (Artursson and Karlsson 1991). Thus, they are expected to
be much less predictive for exposure routes (dermal and inhalation) that are more relevant to acute battlefield exposure. Likewise, crossing the bloodbrain barrier is especially relevant for neurotoxicity, and although there are computational approaches for predicting bloodbrain barrier penetration (Gerebtzoff and Seelig 2006; Carpenter et al.
2014), no in vitro assay accurately measures this property.
8
34
Prepublication Copy

known toxic effects and incorporate more metabolically competent test systems into its test battery. Chapter 4 describes each approach in some detail.
IN SILICO APPROACHES FOR PREDICTING TOXIC EFFECTS
In silico models incorporate a variety of physicochemical features that can be used to predict receptor binding, toxicity, and other biological outcomes. Many (Q)SAR models developed for use in toxicology have been built on a longstanding recognition that the physicochemical properties of a chemical, especially lipophilicity, are highly relevant to prediction of acute toxicity. It has been shown that the presence
or absence of various physicochemical properties can be used to group chemicals into toxicity categories
(Greene and Song 2011). That concept is well established and used in the pharmaceutical industry to
reduce attrition in drug discovery, reduce toxicity, and improve the drug-likeness of chemicals (see Table
3-1).9 Indeed, several studies have shown how simple measures, such as logKow and total polar surface
area (TPSA), provide useful indicators of potential toxicity in vivo. Hughes et al. (2008) showed for a
dataset of 245 substances that substances that had low logKow and high TPSA were about 2.5 times more
likely to be clean (nontoxic) than to be toxic. Precisely the reverse was true of chemicals that had high
logKow and low TPSA, properties that increased the likelihood of chemical binding to multiple biological
targets that could contribute to toxicity.
Although the chemical domain of concern for DOD goes beyond that of drug-like substances, an
understanding of the type of physicochemical properties that can affect adverse outcomes and the range of
property values for which the effect is likely to be substantial can offer useful insights for guiding the assessment of acute toxicity of chemicals of interest to DOD. The sections that follow describe in silico approaches that are available or could be developed for predicting acute toxicity that is relevant to DODs
concerns.
Acute Oral Toxicity
There are few (Q)SAR models and expert systems for prediction of in vivo acute toxicity. The predictiveness of the models, however, can be variable. For some chemicals, the models provide predicted
values that deviate by several orders of magnitude from the experimental data. Furthermore, efforts have
focused largely on the prediction of acute rodent oral toxicity (see Appendix B for a description of various toxicity data sources). Fewer attempts have been made to derive models of acute toxicity via other
routes of exposure, such as dermal or inhalation, although predictions based on extrapolation from acute
oral LD50 values have been attempted.
Available (Q)SARs for acute systemic toxicity have been reviewed (Cronin and Dearden 1995; Cronin et al. 2003; Lessigiarska et al. 2005; Tsakovska et al. 2006; Devillers and Devillers 2009; Lapenna et
al. 2010). Several QSAR models have identified hydrophobicity and electronic and steric effects as important model parameters. Many literature-based models were developed for a single chemical class, such
as alcohols, barbiturates, pyrines and their derivatives, and benzene derivatives. Examples are provided in
Table 3-2.
In contrast, models based on heterogeneous datasets have typically been incorporated into expert
systems. There are, however, examples of models based on heterogeneous data that have not been incorporated into expert systems, and there are examples of models that use a hybrid approach. Table 3-3 provides several examples of various types of models and tools. There has been an evolution in the types of
(Q)SAR models developed over the years to predict acute toxicity. Expert systems tended to favor large
datasets (global models) that use chemistry-based descriptors to derive estimates of rodent oral toxicity.
9
Drug-likeness refers to molecules that contain functional groups or have physical properties similar to those of
known drugs (Walters and Murcko 2002).
Prepublication Copy
35
Hybrid expert systems consider biological activity, such as cytotoxicity information as described in Chapter 4, and chemistry-based descriptors as inputs. More recently, there has been a return to local (Q)SAR
models; they are integrated into batteries of (Q)SARs that can predict acute toxicity of diverse chemicals.
Acute Dermal Toxicity
To the committees knowledge, there are no notable QSARs for the prediction of rodent dermal
acute-toxicity values. Dermal LD50 values might be estimated by extrapolating from oral LD50 values in
some cases by using toxicokinetic information. A case study of three cosmetic substances was performed
by Gajewska et al. (2014) to evaluate such an extrapolation.
Moore et al. (2013) found that the toxicity of chemicals was usually greater by the oral route than
the dermal route. They proposed that data on oral acute systemic toxicity could be used in lieu of equivalent dermal testing with little or no concern for underclassification according to the Globally Harmonized
System of Classification and Labeling of Chemicals (GHS).10 For example, dermal testing of a substance
that has an oral LD50 of greater than 2,000 mg/kg would provide no added value for categorizing its hazard. Moore et al. (2013), however, reported that a majority of chemicals that they evaluated would be
classified more stringently if oral classifications were applied directly to the dermal route. One approach
to address the tendency for overclassification would be to consider whether a chemical is absorbed by the
skin.
Acute Inhalation Toxicity
There are only a handful of QSARs for acute inhalation toxicity. One example is that for volatile
substances. Veith et al. (2009) derived a baseline narcosis model that related vapor pressure (as logVP in
millimeters of mercury) to the 4-hour molar logLC50 in rodents for neutral organic substances. Veith and
Wallace (2006) also established relationships for reactive (electrophilic) chemicals in which reactivity, as
measured in the glutathione depletion assay (Schultz et al. 2005), was related to the molar logLC50. The
underlying basis of their strategy mimics the Adverse Outcome Pathway (AOP) construct as described by
Ankley et al. (2010).
TABLE 3-2 Examples of (Q)SARs for Various Chemical Classes
Chemical Class
Alcohols
Description
Four molecular-structure descriptors and two indicator variables
formed the basis of a categorical model that categorized 95 alcohols
into ranges of LD50 values.
Reference
Guilian and Naibin 1998
Barbiturates
The number of valence electrons and logKow were found to be

predictive of LD50 values of a set of 11 barbiturates.
Hansch and Kurup 2003
Pyrines and derivatives
The energy of the lowest unoccupied molecular orbital and logKow

represented the descriptors used in a QSAR for pyrines and their
derivatives.
Cronin et al. 2002
Benzene derivatives
Electronegativity, dipole moment, and the presence of nitrogen-containing

groups were most important in predicting the acute oral toxicity of benzene
derivatives.
Toropov et al. 2008
10
The GHS is an internationally agreed-on system created by the UN to replace the various classification and labeling standards used in different countries by using consistent criteria for classification and labeling (UNECE
2015).
36
Prepublication Copy
Model
Description
Models Based on Heterogeneous Datasets That Have Been Incorporated into Expert Systems
Outputs
Reference
TOxicity Prediction by KomputerAssisted Technology (TOPKAT)
19 QSAR regression models are based on a number of structural, topological, and

electropological indexes and experimental values for 4,000 chemicals in the
Registry of Toxic Effects of Chemical Substances (RTECS) database.
Rat LD50 (oral)

Rat LC50 (inhalation)
TOPKAT, Accelrys (reviewed

in Lapenna et al. 2010)
Toxicity Estimation Software

Tool (TEST)
Based on chemicals in the RTECS database. Uses a variety of QSAR models

(hierarchical method, Food and Drug Administration [FDA] method, single-model
method, group-contribution method, nearest-neighbor method, and consensus
method) to yield toxicity estimates.
Rat LD50 (oral)
EPA 2014
ACD/Labs Tox suite
Predictions are based on a combination of expert knowledge of various basal and

extracellular effects (such as cholinesterase inhibition, ATP synthesis, and CNS
disruption) and QSAR analysis of more than 10,000 chemicals. Predictions are
provided with reliability estimations, and chemicals are classified into five toxicity
categories.
Rat and mouse LD50 (routes include

oral, intraperitoneal, subcutaneous,
and intravenous)
ACD/Labs 2015
Prediction is based on the analysis of the similarity of chemicals that have known
median LD50 that are taken from a dataset of 38,000 chemicals and incorporates the
identification of toxic fragments.
Models Based on Heterogeneous Datasets That Have Not Been Incorporated into Expert Systems
Rodent LD50 (oral)
Drwal et al. (2014);

ProTox (2015)
Consensus models
Use rodent in vivo acute oral data from the National Library of Medicine databases
as reported in ChemIDplus. Predictions are based on different QSAR statistical
techniques, including random forest, FDA MDL-QSAR programs approach to
k-nearest neighbor, and hierarchical clustering.
Rodent LD50 (oral)
Zhu et al. (2009a)
Multiclassification methods
Based on a dataset containing 12,204 diverse chemicals and published acute oral
Rat LD50 (oral)
rodent LD50s. Model predictions obtained with machine-learning methods, such as
support vector machine, C4.5 decision tree, random forest, k-nearest neighbor, naive
Bayes algorithms, and MACCS and FP4 fingerprints.
Li et al. (2014)
Tiered approach
All chemicals separated into two groups: one based on the relationship between the
in vitro half-maximal inhibitory concentration (IC50) and rodent LD50 and the other
contained the remaining chemicals. A two-step QSAR modeling approach was then
applied. The derived binary classification QSAR models predicted group
membership on the basis of the in vitroin vivo relationships, and a second QSAR
model estimated the LD50s for the chemical subsets.
Rodent LD50 (oral)
Zhu et al. (2009b)
Chemical and biological descriptors
Dataset consisted of 67 chemicals obtained from the literature. Used structural

information and in vitro basal cytotoxicity to predict human acute toxicity.
Indirect measures of human toxicity

(e.g., LC50)
Lee et al. (2010)
Inclusion of concentrationresponse
data derived from high-throughput
screening
Used quantitative high-throughput screening concentrationresponse data to

complement traditional chemical descriptors in the modeling of acute oral
rodent LD50s.
Rodent LD50 (oral)
Sedykh et al. (2011)
ProTox
Prepublication Copy
TABLE 3-3 Examples of Models and Tools for Predicting Acute Oral Toxicity
Global Hybrid Approachesa
(Continued)
37
38
TABLE 3-3 Continued

Model
Local Hybrid Approachesb
Description
Outputs
OASIS Pipeline
Relies on a baseline model for neutral organic substances supplemented with

mechanistic SARs for different reaction-chemistry domains. The approach is
underpinned by 3-D QSARs where relevant to predict acute oral-toxicity categories
by using the same RTECS dataset as used by Zhu et al. (2009a). Complements the
approach outlined by Koleva et al. (2011).
Local lazy method
Based on a dataset of 9,617 chemicals. Uses local structuretoxicity relationships

associated with a query substance to develop acute oral LD50 models.
Reference
Mekenyan et al. (personal
communication, December 2014)
Rodent LD50 (oral)
Lu et al. (2014)
Global hybrid approaches use models derived on the basis of a large heterogeneous dataset that includes chemical and biological information.
Local hybrid approaches aim to derive models that are specific to chemical class or reaction chemistry but will still be applicable to a broad spectrum of chemicals.
b
Prepublication Copy

The expert system TOPKAT incorporates a global model for the prediction of acute inhalation toxicity. The rat inhalation LC50 module contains five models related to different chemical classes to cover a
reasonable breadth of chemical coverage.
Neurotoxicity
Neurotoxicity is another debilitating effect associated with some acute exposures. A few QSARs
have been derived for neurotoxicity, but they are quite limited in scope. Cronin (1996) derived a neurotoxicity QSAR for a set of 44 common solvents that depended on logKow and membrane permeability. A
number of modeling approaches to derive QSAR models for organophosphorus pesticides have also been
developed (Devillers 2004; Garcia-Domenech et al. 2007).
A handful of SARs exist that might identify structural alerts for neurotoxic potential. Chemical classes associated with neurotoxicity include some organic solvents, organophosphorus chemicals, and carbamates, which can induce chronic toxic encephalopathy, delayed neurotoxicity, and cholinergic effects,
respectively. For example, Derek Nexus includes the following structural alerts: organophosphate (for
direct and indirect anticholinesterase activity), N-methyl or N,N-dimethyl carbamate (for direct anticholinesterase activity), and gamma-diketones (for neurotoxicity) (ECHA 2014).
Neurotoxicity is clearly an effect whose mechanisms need to be better elucidated, and models developed accordingly.
Cytotoxicity
Ekwall (1983) suggested that for most chemicals, toxicity was a consequence of nonspecific alterations in cellular function; thus, evaluating the cytotoxic potential of chemicals with cytotoxicity assays
could provide an indication of their potential in vivo toxicity. There have been many attempts to explore
the correlation between in vivo acute toxicity and cytotoxicity data and a number of efforts to predict cytotoxicity from chemical structure. For example, as part of the Multicenter Evaluation of In Vitro Cytotoxicity program (Ekwall et al. 1998), 50 reference chemicals were tested in 61 cytotoxicity assays in the
hope of predicting acute oral LD50s. The coefficients of determination (r2) were 0.61 for rat LD50s and
0.65 for mouse LD50s. And, Lessigiarska et al. (2006) demonstrated how acute toxicity in rats, mice, and
humans could be predicted by using QSAR models that incorporated cytotoxicity data, other biological
end points, and chemical structural descriptor data.
A host of QSAR models have been derived to predict cytotoxicity of various chemical classes. In
many cases, hydrophobicity was a predominant descriptor related to cytotoxicity. For example, McKarns
et al. (1997) correlated the loss of membrane integrity in rat liver epithelial cells with hydrophobicity as
modeled by logKow for a series of 11 alcohols. Other QSARs, as summarized by Tsakovska et al. (2006),
have been developed for p-substituted benzyl alcohols, phenols, anilines, chlorophenols, and polybrominated diphenyl ethers. Papa et al. (2009) developed QSARs for three toxicological end points: mouse oral
LD50 values, inhibition of NADH oxidase (EC50), and effect on mitochondrial membrane potential (EC50).
Freidig et al. (2007) found that nonspecific cytotoxicity could help to identify irritant chemicals.
As part of the European Union framework program, ACuteTox, many investigations were performed to explore in vivoin vitro relationships. Clothier et al. (2013) used Spearman rank-correlation
analysis and hierarchical-cluster analysis to identify in vitro testing strategies for predicting acute toxicity.
Classification-based and regression-based quantitative structuretoxicity relationship (QSTR) and toxicophore models were developed by Kar and Roy (2013). They used in vitro cytotoxicity data collected from
the ACuteTox database.11 Their QSTR models showed that cytotoxicity was influenced by the presence of
hydrophobic aliphatic groups, a ring aromatic group, and hydrogen-bond donors. The in silico models
11
See http://www.acutetox.eu/.
Prepublication Copy
39
derived were considered capable of identifying the essential structural attributes and quantifying the molecular properties that drive in vitro basal cytotoxicity. Prieto et al. (2013) proposed a heuristic testing
strategy for identifying potential neurotoxicants that considered octanolwater partition coefficients, the
prediction results from the neutral red uptake assay performed in 3T3 cells, and in silico predictions of
intestinal absorption and bloodbrain barrier passage.
Limitations and Need for Improvement
The greatest focus in the literature has been on deriving QSAR models to predict oral rodent LD50s.
There are some models for specific chemical classes, but there has been greater interest in exploring the
feasibility of deriving global models. Many of the global models have been data-driven, although some
attempts have included consideration of a chemical mechanistic approach akin to that described for acute
fish toxicity (Bradbury et al. 1990; Schultz et al. 2006). More recently and in part as stimulated by work
within the ACuteTox program, predicting in vivo acute toxicity has considered the use of (Q)SAR approaches in conjunction with in vitro cytotoxicity data. This shift of integrating in vitro and in silico approaches is consistent with the framework of AOPs as one means of incorporating more mechanistic information in testing and assessment approaches for different purposes.
A key issue in all approaches is their relevance and applicability to DOD chemicals. The relevance
of the (Q)SAR models cited earlier would need to be probed for the types of chemicals under consideration by DOD to determine the extent to which the existing models are appropriate in light of the applicability domain and the decision context in question. If the substances of interest are entirely or mostly out
of the applicability domain, new data might need to be identified or other primary sources exploited to
collate and compile more relevant information for the derivation of new models or refinement of existing
models. It will also be critical in such an evaluation to consider the robustness of the training set and associated data that are used to develop the (Q)SAR model. The OECD principles for the validation of
(Q)SAR provide a convenient framework for assessing the validity and applicability of (Q)SAR models
(OECD 2007b).
A second issue is the nonavailability of tools to assess toxicity by nonoral routes of exposure. As
discussed above, most tools have been developed to evaluate the oral exposure route.
A third issue is that the existing (Q)SARs are often lacking in mechanistic basis. Future (Q)SARs
could conceivably be derived to predict key events as reflected in Figure 3-1. (Q)SARs would be developed to estimate outcomes of initial or intermediate events within a pathway rather than to predict an adverse outcome directly as indicated in Figure 3-1. An example of that approach is the early work of the
ACuteTox program in which QSARs were derived to estimate in vitro cytotoxicity rather than in vivo
acute rat toxicity.
Speciation
Parent
Chemical
Initial
Molecular
Events
and
Measurable
Key Events
Adverse
Outcome
Metabolism
(Q)SAR
(Q)SAR
Chemistry or
Biochemistry
Systems
Biology
FIGURE 3-1 Conceptual framework for the future development of (Q)SARs.
40
Prepublication Copy

Another interesting example of exploiting the framework shown in Figure 3-1 focuses on mitochondrial inhibition, a mechanism known to drive acute toxicity of some chemicals. Bhhatari et al. (2014)
compared data from high-throughput screening assays of mitochondrial toxicity in HepG2 cells (see Attene-Ramos et al. 2015) with in silico data on absorption and first-pass metabolism12 and obtained promising results for predicting acute toxicity in multiple species. They found that mitochondrial inhibition
predicted the minimum toxicity in fish and daphnia and that the lower the assay AC50,13 the more likely
that the toxicity was driven by mitochondrial toxicity. However, mitochondrial inhibition did not often
predict the toxicity of chemicals in rats because of the lack of data on oral bioavailability and first-pass
metabolism. However, simulations that used in silico models for bioavailability and metabolism did improve toxicity predictions. A similar approach has been put forward by LHASA Ltd as a contribution to
the OECD AOP work program (OECD 2011). A set of SARs has been derived from substances that inhibit complexes I, III, IV, and V of the electron transport chain, which characterize molecular initiating
events in the pathway that leads to mitochondrial toxicity.
Another route by which models could be derived would involve integrating data from a variety of
inputs (such as in vitro IC50s or AC50s and rodent LD50s) to predict acute toxicity. Several such examples
have been investigated as part of the ACuteTox program as described above. Clothier et al. (2013) used
Spearman rank-correlation analysis and hierarchical clustering to help to identify a combination of in
vitro test systems for predicting in vivo acute toxicity. Kinsner-Ovaskainen et al. (2013) used classification and regression-tree analysis of in vitro data from the ACuteTox program to predict acute oral-toxicity
categories. Kopp-Schneider et al. (2013) likewise investigated various data-mining approaches with the
ACuteTox program data.
A second alternative route of developing (Q)SARs would consider the biological pathways involved
in acute debilitating toxicity (such as altered oxygen transport, changes in neurotransmitter function, and
disruption of cytoskeleton), which could be elucidated in a construct based on mechanistic pathways, and
appropriate assays could be mapped to the key biological events. This approach would provide a different
basis for development of integrated approaches, including specific (Q)SAR models that address specific
key biological events of the AOP. The chemical applicability domain of the assays that characterize each
key event could be extracted to inform new SARs that would facilitate profiling of untested substances.
This type of approach has been attempted for skin sensitization (Patlewicz et al. 2014). Before such a
strategy can be used, an approach to assessing the validity of the assays (Chapter 4), of the prediction
models (data integration models) (Chapter 5), and of the pathways (Chapter 2) needs to be established
(see Patlewicz et al. 2015).
ENSURING SCIENTIFIC CONFIDENCE IN (Q)SAR MODELS
Ensuring scientific confidence in a (Q)SAR model relies on an assessment of model validity and
model applicability. Both are critical for the appropriate interpretation and use of the predictions derived.
The OECD validation principles for (Q)SARs provide a useful construct for evaluating and characterizing
a given (Q)SAR model to determine its scientific validity. The five principles describe the need for a defined end point, an unambiguous algorithm, a defined domain of applicability, measures of performance,
and a mechanistic interpretation if possible (OECD 2004, 2007b). The applicability domain,14 which involves extracting an applicability domain on the basis of the training set used to derive the QSAR model,
provides a basis for judging the relevance and reliability of a prediction made for a given target substance.
There are many ways to extract an applicability domain. Typical approaches include structural, mechanis12
First-pass metabolism can occur at the site of chemical absorption (for example, in the gastrointestinal tract) or
in the liver. In general, first-pass metabolism reduces the amount of a chemical that reaches the systemic circulation.
13
AC50 is the concentration required to elicit a 50% response in an in vitro assay.
14
The applicability domain is the array of chemicals for which the (Q)SAR can confidently be applied for purposes of toxicity prediction (Aptula and Roberts 2006).
Prepublication Copy
41
Appliccation of Mod
dern Toxicology Approachees for Predictting Acute Tooxicity for Cheemical Defennse
bolic, and desscriptor consid
derations (Neetzeva et al. 22005; Dimitroov et al. 20055). Freely avaailable
tic, metab
and comm
mercial tools also exist to
o determine an
a applicabilitty domain, nnamely, AMB
BIT Discoverry and
Domain Manager
M
(Nik
kolova-Jeliazzkova and Jaaworska 20055; Patlewicz et al. 2011).. Sazonovas et al.
(2010) presented an allternative app
proach to extrracting an appplicability doomain for an LD50 model. Each
prediction
n was associaated with a reeliability indeex that dependded on the taargets similaarity to the traaining
set and th
he consistency
y of experim
mental results with regard to the baselinne model in the local cheemical
environment.
The applicability domain form
ms only one facet
f
in judgiing the relevaance and reliaability of a prediction. To ensure
e
the reliability of a prediction,
p
on
ne also needs to evaluate hhow well the models make correct predictions for sim
milar chemicals. Such simillar chemicalss might be ideentified on thee basis of thee same
characteriistics or descrriptors that were
w
used to derive
d
the oriiginal QSAR model or on the basis of strucm a second fa
tural similarity. The prredictivity of similar analo
ogues will form
facet of judginng the relevannce of
a model for
f the chemiical of interesst. The frameework outlineed in the REA
ACH guidancce (see Figuree 3-2)
might be helpful in su
ummarizing th
he key consid
derations for the assessmeent of a (Q)S
SAR model aand its
associated
d prediction.
FIND
DINGS AND RECOMME
ENDATION
NS
Finding:
F
Multtiple databasees are availab
ble for perforrming an initiial chemical characterizatiion of
molecularr structure, ph
hysicochemiccal properties, and availab le acute toxiccity data. In aaddition, a nuumber
of in silico
o models are available for predicting ph
hysicochemiccal properties..
Finding:
F
Physsicochemical and structuraal properties are critical ffor chemical characterizatiion in
that they can help to predict
p
a chem
micals potenttial to pose a physical hazzard, its reacttivity, and its pharmacokineetic characteristics, such as bioavailabiliity and likely routes of expposure.
Finding:
F
Alth
hough a numb
ber of tools are
a available to predict thee site of metaabolism and likely
metabolicc products on the basis of chemical stru
ucture and phhysicochemical properties,, they are dessigned
largely fo
or pharmaceuttical agents. Information
I
about
a
the likeely metabolic products willl be critical ffor informing experimental
e
design,
d
includ
ding assay sellection.
Finding:
F
Seveeral (Q)SAR models
m
that use
u structural properties or physicochem
mical properties are
available for predicting
g acute oral LD
L 50s. Few models
m
for prredicting inhaalation LC50s are availablee, and
none for predicting
p
derrmal LD50s was
w identified.
Finding:
F
A few
w QSAR mod
dels for prediicting neurotooxicity and cyytotoxicity aree available, but not
for other end
e points thaat are relevan
nt for acute, deebilitating toxxicity. Currennt research in (Q)SAR moddels is
focusing on
o incorporatting more bio
ological inform
mation, such as integratingg in vitro datta (for exampple, on
cytotoxiciity) and inform
mation on speecific AOPs.
ments associateed with evaluaating the adequuacy of a (Q))SAR model aand its predicttion as
FIGURE 3-2 Key elem
om the REACH
H technical guiidance.
adapted fro
42
Prrepublication Copy

Recommendation: DOD should evaluate the applicability, relevance, and reliability of available
(Q)SAR models to meet its needs of assessing a chemicals potential to cause acute, debilitating toxicity.
OECD and REACH principles and guidelines for evaluating and characterizing (Q)SAR models might
provide useful frameworks for conducting such an evaluation. Furthermore, DOD should evaluate the
applicability, relevance, and reliability of models and tools for predicting physicochemical properties and
metabolism.
Recommendation: To fill remaining gaps in nontesting approaches, DOD should consider a number of options for further research and development, including extrapolation of oral LD50 to other exposure
routes through pharmacokinetic models; development of new (Q)SAR models for acute lethality, focusing
particularly on inhalation and dermal exposure; and development of (Q)SAR models augmented with biological information, such as in vitro data and information on targets or mechanisms of acute toxicity.
REFERENCES
ACD/Labs. 2015. Acute Toxicity Prediction Module. Data Sheet [online]. Available: http://www.acdlabs.com/down
load/docs/datasheets/datasheet_acute.pdf [accessed March 24, 2015].
AltTox. 2014. Table of Validated and Accepted Alternative Methods [online]. Available: http://alttox.org/mapp/
table-of-validated-and-accepted-alternative-methods/ [accessed March 18, 2015].
Ankley, G.T., R.S. Bennett, R.J. Erickson, D.J. Hoff, M.W. Hornung, R.D. Johnson, D.R. Mount, J.W. Nichols,
C.L. Russom, P.K. Schmieder, J.A. Serrrano, J.E. Tietge, and D.L. Villeneuve. 2010. Adverse outcome pathways: A conceptual framework to support ecotoxicology research and risk assessment. Environ. Toxicol.
Chem. 29(3):730-741.
Aptula, A.O., and D.W. Roberts. 2006. Mechanistic applicability domains for nonanimal-based prediction of toxicological end points: General principles and application to reactive toxicity. Chem. Res. Toxicol. 19(8):10971105.
Artursson, P., and J. Karlsson. 1991. Correlation between oral drug absorption in humans and apparent drug permeability coefficients in human intestinal epithelial (Caco-2) cells. Biochem. Biophys. Res. Commun.
175(3):880-885.
Attene-Ramos, M.S., R. Huang, S. Michael, K.L. Witt, A. Richard, R.R. Tice, A. Simeonov, C.P. Austin, and M.
Xia. 2015. Profiling of the Tox21 chemical collection for mitochondrial function to identify compounds that
acutely decrease mitochondrial membrane potential. Environ. Health Perspect. 123(1):49-56.
Barratt, M.D. 1996. Quantitative structure-activity relationships for skin irritation and corrosivity of neutral and
electrophilic organic chemicals. Toxicol. In Vitro 10(3):247-256.
Bhhatari, B., D. Wilson, P.S. Price, M.J. Bartels, S. Chaunhuri, and E.W. Carney. 2014. Predicting Acute Toxicity
Using In Vitro ToxCast HTS Mitochondrial Inhibition Assays. Presented at the Second ToxCast Data Summit,
September 29-30, 2014, Durham, NC.
Bradbury, S.P., T.R. Henry, and R.W. Carlson. 1990. Fish acute toxicity syndromes in the development of mechanism specific QSARs. Pp. 295-315 in Practical Applications of Quantitative Structure Activity Relationships
(QSAR) in Environmental Chemistry and Toxicology, W. Karcher, and J. Devillers, eds. Dordrecht, The
Netherlands: Kluwer.
Brown, J.S., T. Gordon, O. Price, and B. Asgharian. 2013. Thoracic and respirable particle definitions for human
health risk assessment. Part. Fibre Toxicol. 10:12.
Carpenter, T.S., D.A. Kirshner, E.Y. Lau, S.E. Wong, J.P. Nilmeier, and F.C. Lightstone. 2014. A method to predict
blood-brain barrier permeability of drug-like compounds using molecular dynamics simulations. Biophys. J.
107(3):630-641.
Castillo-Garit, J.A., Y. Marrero-Ponce, F. Torrens, and R. Garca-Domenech. 2008. Estimation of ADME properties
in drug discovery: Predicting Caco-2 cell permeability using atom-based stochastic and non-stochastic linear
indices. J. Pharm. Sci. 97(5):1946-1976.
Clothier, R., M.J. Gmez-Lechn, A. Kinsner-Ovaskainen, A. Kopp-Schneider, J.E. O'Connor, P. Prieto, and S.
Stanzel. 2013. Comparative analysis of eight cytotoxicity assays evaluated within the ACuteTox Project. Toxicol. In Vitro 27(4):1347-1356.
Cronin, M.T. 1996. Quantitative structure-activity relationship (QSAR) analysis of the acute sublethal neurotoxicity
of solvents. Toxicol. In Vitro 10(2):103-110.
Prepublication Copy
43
Cronin, M.T., and J.C. Madden, eds. 2010. In Silico Toxicology: Principles and Applications. London: Royal Society of Chemistry.
Cronin, M.T., and J.C. Dearden.1995. QSAR in toxicology. 2. Prediction of acute mammalian toxicity and interspecies correlations. Quant. Struct. Act. Relat. 14(2):117-120.
Cronin, M.T., J.C. Dearden, J.C. Duff, R. Edwards, N. Manga, A.P. Worth, and A.D. Worgan. 2002. The importance of hydrophobicity and electrophilicity descriptors in mechanistically based QSARs for toxicological
endpoints. SAR QSAR Environ. Res. 13(1):167-176.
Cronin, M.T., J.C. Dearden, J.D. Walker, and A.P. Worth. 2003. Quantitative structure activity relationships for
human health effects: Commonalities with other endpoints. Environ. Toxicol. Chem. 22(8):1829-1843.
Dancik, Y., M.A. Miller, J. Jaworska, and G.B. Kasting. 2013a. Design and performance of a spreadsheet-based
model for estimating bioavailability of chemicals from dermal exposure. Adv. Drug Deliv. Rev. 65(2):221236.
Dancik, Y., J.A. Troutman, and J. Jaworska. 2013b. A framework incorporating the impact of exposure scenarios
and application conditions on risk assessment of chemicals applied to skin. In Silico Pharmacol. 1:10; doi:
10.1186/2193-9616-1-10.
Dearden, J., and A. Worth. 2007. In Silico Prediction of Physicochemical Properties. JRC Scientific and Technical
Reports EUR 23051[online]. Available: https://eurl-ecvam.jrc.ec.europa.eu/laboratories-research/predictive_
toxicology/doc/EUR_23051_EN.pdf [accessed March 16, 2015].
Devillers, J. 2004. Prediction of mammalian toxicity of organophosphorus pesticides from QSTR modeling. SAR
QSAR Environ. Res. 15(5-6):501-510.
Devillers, J., and H. Devillers. 2009. Prediction of acute mammalian toxicity from QSARs and interspecies correlation. SAR QSAR Environ. Res. 20(5-6):467-500.
Dimitrov, S., G. Dimitrova, T. Pavlov, N. Dimitrova, G. Patlewicz, J. Niemela, and O. Mekenyan. 2005. A stepwise
approach for defining the applicability domain of SAR and QSAR models. J. Chem. Inf. Model. 45(4):839-849.
Drwal, M.N., P. Banerjee, M. Dunkel, M.R. Wettig, and R. Preissner. 2014. ProTox: A web server for the in silico
prediction of rodent oral toxicity. Nucleic Acids Res. 42(Web Server issue):W53-W58.
ECETOC (European Centre for Ecotoxicology and Toxicology of Chemicals), 2012. Category Approaches, Readacross, (Q)SAR. Technical Report No. 116. Brussels, Belgium: ECETOC.
ECHA (European Chemicals Agency). 2008. Guidance on Information Requirements and Chemical Safety Assessment. Chapter R.6: QSARs and Grouping of Chemicals [online]. Available: http://echa.europa.eu/documents/
10162/13632/information_requirements_r6_en.pdf [accessed February 12, 2015].
ECHA (European Chemical Agency). 2012. Guidance on Registration, Version 2.0. Guidance for the Implementation of REACH. May 2012 [online]. Available: http://echa.europa.eu/documents/10162/13632/registration_
en.pdf [accessed April 30, 2015].
ECHA 2014. Guidance on Information Requirements and Chemical Safety Assessment, Chapter R.7a. Endpoint Specific Guidance, Version 3.0 [online]. Available: http://echa.europa.eu/documents/10162/13632/information_re
quirements_r7a_en.pdf.
Ekwall, B. 1983. Screening of toxic compounds in mammalian cell cultures. Ann. N.Y. Acad. Sci. 407: 64-77.
Ekwall, B., F.A. Barile, A. Castano, C. Clemedson, R.H. Clothier, P. Dierickx, B. Ekwall, M. Ferro, G. Fiskesj, L.
Garza-Ocanas, M.J. Gmez-Lechn, M. Glden, T. Hall, B. Isomaa, A. Kahru, G. Kerszman, U. Kristen, M.
Kunimoto, S. Krenlampi, L. Lewan, A. Loukianov, T. Ohno, G. Persoone, L. Romert, T.W. Sawyer, H. Segner, R. Shrivastava, A. Stammati, N. Tanaka, M. Valentino, E. Walum, and F. Zucco. 1998. MEIC evaluation
of acute systemic toxicity. Part VI. Prediction of human toxicity by rodent LD50 values and results from 61 in
vitro tests. ATLA 26(Suppl. 2):617-658.
EPA (US Environmental Protection Agency). 2012. Glossary of Terms: Methods of Toxicity Testing and Risk Assessment [online]. Available: http://www.epa.gov/pesticides/science/comptox-glossary.html#q [accessed May7,
2015].
EPA (US Environmental protection Agency). 2014. Toxicity Estimation Software Tool (TEST)[online]. Available:
http://www.epa.gov/nrmrl/std/qsar/qsar.html#TEST [accessed March 18, 2015].
EU (European Union). 2006. Regulation (EC) No. 1907/2006 of the European Parliament and of the Council concerning the Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH), establishing a European Chemicals Agency, amending Directive 1999/45/EC and repealing Council Regulation (EEC) No.
793/93 and Commission Regulation (EC) No. 1488/94 as well as Council Directive 76/769/EEC and Commission Directives 91/155/EEC, 93/67/EEC, 93/105/EC and 2000/21/EC..OJ E.U. 396:I1-I849 [online]. Available: http://faolex.fao.org/docs/pdf/eur68317.pdf.
44
Prepublication Copy

Fitzpatrick, D., J. Corish, and B. Hayes. 2004. Modelling skin permeability in risk assessment--the future. Chemosphere 55(10):1309-1314.
Flynn. G.L. 1990. Physicochemical determinants of skin absorption. Pp. 93-127 in Principles of Route-to-Route
Extrapolation for Risk Assessment, T.R. Gerrity, and C.J. Henry, eds. New York: Elsevier.
Freidig, A.P., S. Dekkers, M. Verwei, E. Zvinavashe, J.G. Bessems, and J.J. van de Sandt. 2007. Development of a
QSAR for worst case estimates of acute toxicity of chemically reactive compounds. Toxicol. Lett. 170(3):214222.
Gajewska, M., A. Worth, C. Urani, H. Briesen, and K.W. Schramm. 2014 Application of physiologically-based toxicokinetic modelling in oral-to-dermal extrapolation of threshold doses of cosmetic ingredients. Toxicol. Lett.
227(3):189-202.
Garca-Domenech, R., P. Alarcn-Elbal, G. Bolas, R. Bueno-Mar, F.A. Chord-Olmos, S.A. Delacour, M.C. Mourio, A. Vidal, and J. Glvez. 2007. Prediction of acute toxicity of organophosphorus pesticides using topological indices. SAR QSAR Environ. Res. 18(7-8):745-755.
Gerebtzoff, G., and A. Seelig. 2006. In silico prediction of blood-brain barrier permeation using the calculated molecular cross-sectional area as main parameter. J. Chem. Inf. Model. 46(6):2638-2650.
Gerner, I., M.D. Barratt, S. Zinke, K. Schlegel, and E. Schlede. 2004. Development and prevalidation of a list of
structure-activity relationship rules to be used in expert systems for prediction of the skin-sensitizing properties of chemicals. ATLA 32(5):487-509.
Gleeson, M.P. 2008. Generation of a set of simple, interpretable ADMET rules of thumb. J. Med. Chem. 51(4):817834.
Grbic, S., J. Parojcic, S. Ibric, and Z. Djuric. 2011. In vitro-in vivo correlation for gliclazide immediate-release tablets based on mechanistic absorption simulation. AAPS PharmSciTech. 12(1):166-171.
Green, N., and M. Song. 2011. Predicting in vivo safety characteristics using physiochemical properties and in vitro
assays. Fut. Med. Chem. 3(12):1503-1511.
Guerra, A., N.E. Campillo, and J.A. Pez. 2010. Neural computational prediction of oral drug absorption based on
CODES 2D descriptors. Eur. J. Med. Chem. 45(3):930-940.
Guilian, W., and B. Naibin. 1998. Structure activity relationships for rat and mouse LD50 of miscellaneous alcohols.
Chemosphere 36(7):1475-1483.
Hansch, C., and A. Kurup. 2003. QSAR of chemical polarisability and nerve toxicity, 2. J. Chem. Inf. Comp. Sci.
43(5):1647-1651.
Hostynek, J.J., and P.S. Magee. 1997. Modelling in vivo human skin absorption. Molec. Inform. 16(6):473-479.
Hughes, J.D., J. Blagg, D.A. Price, S. Bailey, G.A. DeCrescenzo, R.V. Devraj, E. Ellsworth, Y. Fobian, M.E. Gibbs,
R.W. Gilles, N. Greene, E. Huang, T. Krieger-Burke, T. Wager, L. Whiteley, and Y. Zhang. 2008. Physiochemical drug properties associated with in vivo toxicological outcomes. Biorg. Med. Chem. Lett.
18(17):4872-4875.
Hulzebos, E., J.D. Walker, I. Gerner, and K. Schlegel. 2005. Use of structural alerts to develop rules for identifying
chemical substances with skin irritation or skin corrosion potential. QSAR Comb. Sci. 24(3):332-342.
Iyer, M., Y.J. Tseng, C.L. Senese, J. Liu, and A.J. Hopfinger. 2007. Prediction and mechanistic interpretation of
human oral drug absorption using MI-QSAR analysis. Mol. Pharm. 4(2):218-231.
Kar, S., and K. Roy. 2013. First report on predictive chemometric modeling, 3D-toxicophore mapping and in silico
screening of in vitro basal cytotoxicity of diverse organic chemicals. Toxicol. In Vitro. 27(2):597-608.
Kinsner-Ovaskainen, A., P. Prieto, S. Stanzel, and A. Kopp-Schneider. 2013. Selection of test methods to be included in a testing strategy to predict acute oral toxicity: An approach based on statistical analysis of data collected
in phase 1 of the ACuteTox project. Toxicol. In Vitro 27(4):1377-1394.
Kirchmair, J., M.J. Williamson, J.D. Tyzack, L. Tan, P.J. Bond, A. Bender, and R.C. Glen. 2012. Computational
prediction of metabolism: Sites, products, SAR, P450, enzyme dynamics and mechanism. J. Chem. Inf. Model. 52(3):617-648.
Kirman, C.R., L.L. Aylward, B.A. Wetmore, R.S. Thomas, M. Sochaski, S.S. Ferguson, S.A. Csiszar, and O. Jolliet.
In press. Quantitative property-property relationship for screening-level prediction of intrinsic clearance: A
tool for exposure modeling for high-throughput toxicity screening data. Appl. In Vitro Toxicol.
Koleva, Y.K., M.T. Cronin, J.C. Madden, and J.A. Schwbel. 2011. Modelling acute oral mammalian toxicity. 1.
Definition of a quantifiable baseline effect. Toxicol. In Vitro 25(7):1281-1293.
Kopp-Schneider, A., P. Prieto, A. Kinsner-Ovaskainen, and S. Stanzel. 2013. Design of a testing strategy using nonanimal based test methods: Lessons learnt from the ACuteTox project. Toxicol. In Vitro 27(4):1395-1401.
Prepublication Copy
45
Kuentz, M., S. Nick, N. Parrott, and D. Rthlisberger. 2006. A strategy for preclinical formulation development using GastroPlus as pharmacokinetic simulation tool and a statistical screening design applied to a dog study.
Eur. J. Pharm. Sci. 27(1):91-99.
Langer, T., and G. Wolber. 2004. Pharmacophore definition and 3D searches. Drug Discov. Today Technol.
1(3):203-207.
Lapenna, S., M. Fuart-Gatnik, and A. Worth. 2010. Review of QSAR Models and Software Tools for Predicting
Acute and Chronic Systemic Toxicity. JRC Technical Report EUR 24639 EN. Luxembourg: Office of the European Union.
Lee, S., K. Park, H.S. Ahn, and D. Kim. 2010. Importance of structural information in predicting human acute toxicity from in vitro cytotoxicity data. Toxicol. Appl. Pharmacol. 246(1-2):38-48.
Leeson, P.D., and B. Springthorpe. 2007. The influence of drug-like concepts on decision-making in medicinal
chemistry. Nat. Rev. Drug Discov. 6(11):881-890.
Lessigiarska, I., A.P. Worth, and T.I. Netzeva. 2005. Comparative Review of QSARs for Acute Toxicity. JRC report
EUR 21559 EN. European Commission, Joint Research Centre, Ispra, Italy.
Lessigiarska, I., A.P. Worth, T.I. Netzeva, J.C. Dearden, and M.T. Cronin. 2006. Quantitative structure-activityactivity and quantitative structure-activity investigations of human and rodent toxicity. Chemosphere
65(10):1878-1887.
Li, X., L. Chen, F. Cheng, Z. Wu, H. Bian, C. Xu, W. Li, G. Liu, X. Shen, and Y. Tang. 2014. In silico prediction of
chemical acute oral toxicity using multi-classification methods. J. Chem. Inf. Model. 54(4):1061-1069.
Lipinski, C.A., F. Lombardo, B.W. Dominy, and P.J. Feeney. 2001. Experimental and computational approaches to
estimate solubility and permeability in drug discovery and development settings. Adv. Drug Deliv. Rev. 46(13):3-26.
Lu, J., J. Peng, J. Wang, Q. Shen, Y. Bi, L. Gong, M. Zheng, X. Luo, W. Zhu, H. Jiang, and K. Chen. 2014. Estimation of acute oral toxicity in rat using local lazy learning. J. Cheminform. 6:26; doi:10.1186/1758-2946-6-26.
Magnusson, B.M., Y.G. Anissimov, S.E. Cross, and M.S. Roberts. 2004. Molecular size as the main determinant of
solute maximum flux across the skin. J. Invest. Dermatol. 122(4):993-999.
McKarns, S.C., C. Hansch, W.S. Caldwell, W.T. Morgan, S.K. Moore, and D.J. Doolittle. 1997. Correlation between hydrophobicity of short chain aliphatic alcohols and their ability to alter plasma membrane integrity.
Fundam. Appl. Toxicol. 36(1):62-70.
Meek, M.E., H.A. Barton, J.G. Bessems, J.C. Lipscomb, and K. Krishnan. 2013. Case study illustrating the WHO
IPCS guidance on characterization and application of physiologically based pharmacokinetic models in risk
assessment. Regul. Toxicol. Pharmacol. 66(1):116-129.
Mekenyan, O.G., P.I. Petkov, S.V. Kotov, S. Stoeva, V.B. Kamenska, S.D. Dimitrov, M. Honma, M. Hayashi, R.
Benigni, E.M. Donner, and G. Patlewicz. 2012. Investigating the relationship between in vitro-in vivo genotoxicity: Derivation of mechanistic QSAR models for in vivo liver genotoxicity and in vivo bone marrow micronucleus formation which encompass metabolism. Chem. Res. Toxicol. 25(2):277-296.
Mitragotri, S., Y.G. Anissimov, A.L. Bunge, H.F. Frasch, R.H. Guy, J. Hadgraft, G.B. Kasting, M.E. Lane, and M.S.
Roberts. 2011. Mathematical models of skin permeability: An overview. Int. J. Pharm. 418(1):115-129.
Moore, N.P., D.J. Andrew, D.L. Bjerke, S. Creton, D. Dreher, T. Homes, P. Prieto, T. Seidle, and T.G. Rowan.
2013. Can acute dermal systematic toxicity tests be replaced with oral tests? A comparison of route specific
systemic toxicity and hazard classifications under the Globally Harmonised System of Classification and Labelling of Chemicals (GHS). Regul. Toxicol. Pharmcol. 66(1):30-37.
Netzeva, T., A. Worth, T. Aldenberg, R. Benigni, M. Cronin, P. Gramatica, J. Jaworska, S. Kahn, G. Klopman, C.
Marchant, G. Myatt, N. Nikolova-Jeliazkova, G. Patlewicz, R. Perkins, D. Roberts, T. Schultz, D. Stanton, J.
van de Sandt, W. Tong, G. Veith, and C. Yang. 2005. Current status of methods for defining the applicability
domain of (quantitative) structureactivity relationships. ATLA 33(2):155-173.
Nikolova-Jeliazkova, N., and J. Jaworska. 2005. An approach to determining applicability domains for QSAR group
contribution models: An analysis of SRC KOWWIN. ATLA 33(5):461-470.
NIOSH (National Institute for Occupational Safety and Health)). 2013. Skin Exposures and Effects: Finite Dose
Skin Permeation Calculator [online]. Available: http://www.cdc.gov/niosh/topics/skin/finiteSkinPermCalc.html
NRC (National Research Council). 2014. Physicochemical properties and chemical fate. Pp. 47-67 in A Framework
to Guide Selection of Chemical Alternatives. Washington, DC: National Academies Press.
OECD (Organisation for Economic Co-operation and Development). 2004. OECD Principles for the Validation, For
Regulatory Purposes of (Quantitative) Structure-Activity Relationship Models [online]. Available: http://
www.oecd.org/chemicalsafety/risk-assessment/37849783.pdf [accessed March 17, 2015].
46
Prepublication Copy

OECD (Organisation for Economic Co-operation and Development). 2006. OECD Guideline for Testing of Chemicals No. 435: In Vitro Membrane Barrier Test Method for Skin Corrosion. Adopted July 19, 2006 [online].
Available: http://ntp.niehs.nih.gov/iccvam/suppdocs/feddocs/oecd/oecdtg435.pdf [accessed March 18, 2015].
OECD (Organisation for Economic Co-operation and Development). 2007a. Guidance on Grouping of Chemicals.
OECD Environmental Health and Safety Publication Series on Testing and Assessment No. 80. Paris: OECD
[online]. Available: http://www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?doclanguage=en&co
te=env/jm/mono(2007)28.
OECD (Organisation for Economic Co-operation and Development). 2007b. Guidance Document on the Validation of
(Quantitative) Structure-Activity Relationships [(Q)SAR] Models. OECD Environment Health and Safety Publication Series on Testing and Assessment No. 69. Paris: OECD [online]. Available: http://www.oecd.org/official
documents/publicdisplaydocumentpdf/?doclanguage=en&cote=env/jm/mono(2007)2 [accessed March 17, 2015].
OECD (Organisation for Economic Co-operation and Development). 2011. Report of the Workshop on Using
Mechanistic Information in Forming Chemical Categories, December, 8-10, 2010, Crystal City, VA. OECD
Environment Health and Safety Publication Series on Testing and Assessment No.138. Paris: OECD [online].
Available: http://www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=env/jm/mono(2011)8&do
clanguage=en [accessed March 17, 2015].
OECD (Organisation for Economic Co-operation and Development). 2013a. OECD Guideline for Testing of Chemicals No. 431: In Vitro Skin Corrosion: Reconstructed Human Epidermis (RHE) Test Method. Adopted July
26, 2013 [online]. Available: https://ntp.niehs.nih.gov/iccvam/suppdocs/feddocs/oecd/oecd-tg431-2013-508.pdf
OECD (Organisation for Economic Co-operation and Development). 2013b. OECD Guideline for Testing of Chemicals No. 437: Bovine Corneal Opacity and Permeability Test Method for Identifying i) Chemicals Inducing
Serious Eye Damage and ii) Chemicals Not Requiring Classification for Eye Irritation or Serious Eye. Adopted July 26, 2013 [online]. Available: http://ntp.niehs.nih.gov/iccvam/suppdocs/feddocs/oecd/oecd-tg4372013-508.pdf [accessed March 18, 2015].
OECD (Organisation for Economic Co-operation and Development). 2013c. OECD Guideline for Testing of Chemicals No. 438: Isolated Chicken Eye Test Method for Identifying i) Chemicals Inducing Serious Eye Damage and
ii) Chemicals Not Requiring Classification for Eye Irritation or Serious Eye Damage. Adopted July 26, 2013
[online]. Available: http://ntp.niehs.nih.gov/iccvam/suppdocs/feddocs/oecd/oecd-tg438-2013-508.pdf [accessed
March 18, 2015].
OECD (Organisation for Economic Co-operation and Development). 2014. Guidance on Grouping of Chemicals,
Second Edition. OECD Environment Health and Safety Publication Series on Testing and Assessment No.194.
Paris: OECD [online]. Available: http://www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=env/
jm/mono(2014)4&doclanguage=en [accessed March 17, 2015].
Papa, E., M. Luini, and P. Gramatica. 2009. Quantitative structure-activity relationship modelling of oral acute toxicity and cytotoxic activity of fragrance materials in rodents. SAR QSAR Environ. Res. 20(7-8):767-779.
Patlewicz, G., M.W. Chen, and C.A. Bellin. 2011. Non-testing approaches under REACH--help or hindrance? Perspectives from a practitioner within industry. SAR QSAR Environ. Res. 22(1-2):67-88.
Patlewicz, G., C. Kuseva, A. Kesoba, I. Popova, T. Zhechev, T. Pavlov, D.W. Roberts, and O. Mekenyan. 2014.
Towards AOP application implementation of an integrated approach to testing and assessment (IATA) into a
pipeline tool for skin sensitization. Regul. Toxicol. Pharmacol. 69(3): 529-545.
Patlewicz, G., T. Simon, R. Budinsky, J.C. Rowlands, and R.A. Becker. 2015. Proposing a scientific confidence
framework to help support the application of adverse outcome pathways for regulatory purposes. Regul. Toxicol. Pharmacol. 71(3):463-477.
Perrin, D.D., B. Dempsey, and E.P. Serjeant. 1981. pKa Prediction for Organic Acids and Bases. London: Chapman
and Hall.
Peyret, T., and K. Krishnan. 2012. Quantitative property-property relationship for screening-level prediction of intrinsic clearance of volatile organic chemicals in rats and its integration within PBPK models to predict inhalation pharmacokinetics in humans. J. Toxicol. 2012:286079.
Potts, R.O., and R.H. Guy. 1992. Predicting skin permeability. Pharm. Res. 9(5):663-669.
Prieto, P., A. Kinsner-Ovaskainen, S. Stanzel, B. Albella, P. Artursson, N. Campillo, R. Cecchelli, L. Cerrato, L.
Daz, E. Di Consiglio, A. Guerra, L. Gombau, G. Herrera, P. Honegger, C. Landry, J.E. O'Connor, J.A. Pez,
G. Quintas, R. Svensson, L. Turco, M.G. Zurich, M.J. Zurbano, and A. Kopp-Schneider. 2013. The value of
selected in vitro and in silico methods to predict acute oral toxicity in a regulatory context: Results from the
European Project ACuteTox. Toxicol. In Vitro 27(4):1357-1376.
Prepublication Copy
47
ProTox. 2015. Prediction of Rodent Oral Toxicology. Charite University of Medicine, Berlin, Germany [online].
Available: http://tox.charite.de/tox/ [accessed March 18, 2015].
Raunio, H. 2011. In silico toxicology non-testing methods. Front. Pharmacol. 2:33.
Rorije, E., and E.M. Hulzebos. 2005. Evaluation of (Q)SARs for the Prediction of Skin Irritation/ Corrosion Potential.
National Institute of Public Health and Environment (RIVM-SEC)/European Chemicals Bureau (ECB), Bilthoven, The Netherlands [online]. Available: https://eurl-ecvam.jrc.ec.europa.eu/laboratories-research/predictive_
toxicology/information-sources/qsar-document-area/Evaluation_of_Skin_Irritation_QSARs.pdf [accessed March
17, 2015].
Rostami-Hodjegan, A., and G.T. Tucker. 2007. Simulation and prediction of in vivo drug metabolism in human
populations from in vitro data. Nat. Rev. Drug Discov. 6(2):140-148.
Saliner, A.G., I. Tsakovska, M. Pavan, G. Patlewicz, and A.P. Worth. 2007. Evaluation of SARs for the prediction
of skin irritation/corrosion potential: Structural inclusion rules in the BfR decision support system. SAR
QSAR Environ. Res. 18(3-4):331-342.
Saliner, A.G., G. Patlewicz, and A.P. Worth. 2008. A review of (Q)SAR models for skin and eye irritation and corrosion. QSAR Comb. Sci. 27(1):49-59.
Sazonovas, A., P. Japertas, and R. Diziapetris. 2010. Estimation of reliability of predictions and model applicability
domain evaluation in the analysis of acute toxicity (LD50). SAR QSAR Environ. Res. 21(1):127-148.
Schultz, T.W., J.W. Yarbrough, and E.L. Johnson. 2005. Structure-activity relationships for reactivity of carbonylcontaining compounds with glutathione. SAR QSAR Environ. Res. 16(4):313-322.
Schultz, T.W., R.E. Carlson, M.T. Cronin, J.L. Hermens, R. Johnson, P.J. O'Brien, D.W. Roberts, A. Siraki, K.B.
Wallace, and G.D. Veith. 2006. A conceptual framework for predicting the toxicity of reactive chemicals:
Modeling soft electrophilicity. SAR QSAR Environ. Res. 17(4):413-428.
Sedykh, A., H. Zhu, H. Tang, L. Zhang, A. Richard, I. Rusyn, and A. Tropsha. 2011. Use of in vitro HTS derived
concentration-response data as biological descriptors improves the accuracy of QSAR models of in vivo toxicity. Environ. Health Perspect. 119(3):364-370.
Silva, F.T., and G.H. Trossini. 2014. The survey of the use of QSAR methods to determine intestinal absorption and
oral bioavailability during drug design. Med. Chem. 10(5):441-448.
Simulations Plus. 2010. Manual GastroPlus. Simulations Plus, Inc., CA.
Suenderhauf, C., F. Hammann, A. Maunz, C. Helma, and J. Huwyler. 2011. Combinatorial QSAR modeling of human intestinal absorption. Mol. Pharm. 8(1):213-224.
ten Berge, W. 2009. A simple dermal absorption model: Derivation and application. Chemosphere 75(11):1440-1445.
Toropov, A.A., B.F. Rasulev, and J. Leszcznski. 2008. QSAR Modeling of acute toxicity by balance of correlations.
Bioorg. Med. Chem. 16(11):5999-6008.
Tsakovska, I., I. Lessigiarska, T. Netzeva, and A.P. Worth. 2006. Review of (Q)SARs for Mammalian Toxicity.
EUR 22846 EN. European Commission, Directorate General, Joint Research Centre, Ispra, Italy [online].
Available: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.103.6822&rep=rep1&type=pdf [accessed
March 17, 2015].
Tsakovska, I., A.G. Saliner, T. Netzeva, M. Pavan, and A.P. Worth. 2007. Evaluation of SARs for the prediction of
eye irritation/corrosion potential: Structural inclusion rules in the BfR decision support system. SAR QSAR
Environ. Res. 18(3-4):221-235.
UNECE (U.N. Economic Commission for Europe). 2015. Globally Harmonized System of Classification and Labelling
of Chemicals (GHS) [online]. Available: http://www.unece.org/trans/danger/publi/ghs/ghs_welcome_e.html [accessed March 18, 2015].
Veith, G.D., and K.B. Wallace. 2006. Inhalation Toxicity in Fish and Rodents. Presented at the McKim Conference,
June 27-29, 2006, Duluth, MN.
Veith, G.D., E.P. Petkova, and K.B. Wallace. 2009. A baseline inhalation toxicity model for narcosis in mammals.
SAR QSAR Environ. Res. 20(5-6):567-578.
Walters, W.P., and M.A. Murcko. 2002. Prediction of 'drug-likeness'. Adv. Drug Deliv. Rev. 54(3):255-271.
Wang, J., and T. Hou. In press. Advances in computationally modeling human oral bioavailability. Adv. Drug Deliv.
Rev.
Worth, A. 2008. QSAR and Chemical Read-Across within the Context of Non-testing Strategies. Presentation at the
1st SETAC Europe Special Science Symposium: Integrated Testing Strategies for REACH, October, 23-24,
2008, Brussels, Belgium [online]. Available: http://reach.setac.eu/embed/presentations_reach/09_WORTH_
QSAR_Chem_Read_Across.pdf
Xu, C., and D.E. Mager. 2011. Quantitative structure-pharmacokinetic relationships. Expert Opin. Drug Metab. Toxicol. 7(1):63-77.
48
Prepublication Copy

Yamazaki, K., and M. Kanaoka. 2004. Computational prediction of the plasma protein-binding percent of diverse
pharmaceutical compounds. J. Pharm. Sci. 93(6):1480-1494.
Yang, J., M. Jamei, K.R. Yeo, G.I. Tucker, and A. Rostami-Hodjegan. 2007. Prediction of intestinal first-pass drug
metabolism. Curr. Drug Metab. 8(7):676-684.
Zhu, H., T.M. Martin, L. Ye, A. Sedykh, D. Young, and A. Tropsha. 2009a. Quantitative structure-activity relationship modelling of rate acute toxicity by oral exposure. Chem. Res. Toxicol. 22(12):1913-1921.
Zhu, H., L. Ye, A. Richard, A. Gobraikh, F.A. Wright, I. Rusyn, and A. Tropsha. 2009b. A novel 2-step hierarchical
quantitative structure-activity relationship modelling workflow for predicting acute toxicity of chemicals in
rodents. Environ. Health Perspect. 117(8):1257-1264.
Prepublication Copy
49
4
Assays for Predicting Acute Toxicity
The future path of toxicity testing was foreshadowed early in the 2000s with publication of frameworks or roadmaps that called for an increased emphasis on the use of in vitro assays that evaluate key
biological pathways and molecular mechanisms linked to human disease (EPA 2003; NTP 2004). Highthroughput testing would allow less expensive, rapid screening of large numbers of chemicals to set testing priorities on the basis of predicted adverse health effects. The National Research Council (NRC) report Toxicity Testing in the 21st Century: A Vision and a Strategy (NRC 2007) built on the early publications and gave rise to a variety of large-scale initiatives to see how in vitro testing methods can be used to
predict human toxicity. To implement the strategy outlined in the NRC report, a collaboration was formed
between the National Toxicology Program (NTP), the US Environmental Protection Agency (EPA) National Center for Computational Toxicology, and the National Institutes of Health National Chemical Genomics Center (NCGC)1 to identify mechanisms of chemically induced biological activity, to set priorities
among chemicals for more extensive toxicological evaluation, and to develop more predictive models of
human biological response (MOU 20082; Austin et al. 2008; Kavlock et al. 2009; Krewski et al. 2009).
The collaboration is now formally referred to as the Tox21 program.
The EPA-sponsored ToxCast program (Dix et al. 2007; Kavlock et al. 2012), the Tox21 program,
and the European ACuteToX program (Clemedson et al. 2007; Clemedson 2008) are specific examples of
large-scale initiatives to evaluate in vitro testing methods for their ability to predict human toxicity. Phase
I of the ToxCast program evaluated about 300 conventional pesticide active ingredients in a battery of
cell-free and cell-based assays (Judson et al. 2010). In phase II, the chemical space was broadened to include chemicals used in consumer products and industrial processes and unmarketed drugs donated by
pharmaceutical companies (Kavlock et al. 2012; Sipes et al. 2013). The completed first phase of the
Tox21 screening program tested about 2,800 chemicalsincluding solvents, fire retardants, dyes, preservatives, plasticizers, therapeutic agents, inorganic and organic pollutants, drinking water-disinfection
byproducts, and natural productsin 50 assays (Shukla et al. 2010; Attene-Ramos 2013). The studies laid
the groundwork for efforts to characterize the ability of cell-free and cell-based assays and data-modeling
approaches to predict activity and potency in selected biochemical targets. Most of the assays developed
and validated for high-throughput screening (HTS) applications like the ToxCast program provide information about activation of molecular-receptor families or biochemical activities that are of interest to the
pharmaceutical industry. In fact, for most assays, there is not a direct linkage between specific cells or
tissue and chemical or mechanistic targets associated with acute lethal or debilitating effects outlined in
Table 2-1. Thus, the assays might be of less use for identifying chemicals that potentially can cause acute,
debilitating injuries in deployed personnel.
This chapter reviews currently available tools and their limitations for immediate implementation (in
the next 3-10 years) by the US Department of Defense (DOD) to screen chemicals of interest to DOD for
acute toxicity.
NCGC is now part of the National Center for Advancing Translational Sciences.
High Throughput Screening, Toxicity Pathway Profiling, and Biological Interpretation of Findings, Memorandum of Understanding Between NTP, NCGC and EPA, January 2008. Available: http://www.niehs.nih.gov/about/
highlights/assets/docs/memorandum_of_understanding_508.pdf.
2
50
Prepublication Copy

IN VITRO ASSAYS
Numerous in vitro screening assays have been developed for measuring specific biological activities
of chemicals in specific organs or cell types with an eye to elucidating mechanisms of action. For the purposes of hazard assessment, biological activity in an in vitro system can identify a mechanism of action or
response that could be extrapolated to an in vivo end point. This section reviews relevant in vitro biochemical assays that could potentially be used to assess acute toxicity.
Specific-Protein Assays
Testing whether a chemical inhibits a particular enzyme or binds to a particular receptor or other
biomolecule is the most direct way to test a chemical for a specific mechanism of action at the molecular
level. Enzyme and receptor-binding assays tend to be reliable, to exhibit relatively good agreement between different laboratories, and to be well suited for high-throughput formats. The specific protein, protein complex, receptor, or other biomolecule of interest can be provided in the assay as a pure molecule,
as a partially purified complex, or as a component of living cells.3
Given their high value for predicting specific molecular effects, protein assays are the most frequent
type of assay in the ToxCast program (Figure 4-1). Some of the ToxCast assays have direct relevance to
predicting acute toxicity of chemicals that could be used as warfare agents. For example, assays of acetylcholinesterase activity are applicable to cholinesterase-inhibiting nerve agents, and mitochondrial electron-transport assays are applicable to cyanide. Specific-protein assays in the ToxCast and ACuteTox
studies that were designed to detect nervous system effects, such as modulation of ion-channel activity,
are also relevant for predicting acute toxicity. However, it is important to recognize that, like many of the
other assays in ToxCast, most of the specific-protein assays included in ToxCast were designed for needs
other than predicting acute toxicityfor example, to predict endocrine-disrupting activityand might
have little value for predicting acute toxicity.
Limitations and Needs for Improvement of Protein Assays
A major limitation is the biological space that is covered by specific-protein assays that are designed to assess the actions of a chemical on specific enzymes or receptors. More complicated or nonspecific mechanisms might also be involved in acute toxicity. For example, vesicant action on the skin is not
mediated by the action of a chemical on a specific enzyme or receptor. Toxic chemicals that act through
nonspecific mechanisms might not be identified in specific-protein assays or might be identified in multiple assays but with poor correlation between dose and response, which potentially confuses analyses. Another limitation is that assays for particular acute toxic mechanisms (such as activity on particular ion
channels) might be missing from the suite of assays or might be unreliable. A final limitation is that some
specific-protein assays involve general mechanisms that are common to many cell types and thus might
not predict particular organ toxicities themselves.
To make specific-protein assays more useful for DODs needs than they are now, there is a need
to identify the subset of existing specific-protein assays that are directly mechanistically and quantitatively relevant to debilitating injuries, to fill gaps in assay types (for example, to develop assays for perturbation of particular ion channels that are not included in current platforms), to develop methods for identifying and classifying chemicals that have potent but nonspecific toxic actions, and to evaluate performance
and predictive value of specific-protein assays for predicting acute toxicity by using a panel of reference
and test chemicals relevant to DODs interests.
Assays that use living cells to evaluate the effects of chemicals on specific proteins could have been considered
in the section that follows, but for simplicity, this section reviews all assays whose purpose is to measure effects of a
chemical on the biochemical activity of a specific protein or other biomolecule.
Prepublication Copy
51
Appliccation of Mod
FIGURE 4-1
4 Intended taarget families and subfamiliees for the ToxC
Cast program. The number oof assays for eaach intended targ
get is representted by the sizees and font sizees of the orangge nodes. For eeach intended ttarget family, tthe assays are su
ubdivided into assay
a
end poin
nts (see EPA 20
014a, p.13).
Cell-Based Phenotypic Assays

Thiss section descrribes high-thrroughput assay
ys that use cuultured cells annd measure soome overall pphenotypic outp
put that is releevant to prediccting acute tox
xicity, such ass cell proliferration, plasmaa membrane ppermeability, orr adenosine triphosphate (A
ATP) content.. There is a laarge scientificc literature onn the applicatiion of
cell-based
d assays to dru
ug developmeent and a grow
wing literaturee on their appplication to toxxicology. Thee ToxCast progrram includes more than 10
00 cell-based assays
a
whosee purpose is too measure cyttotoxicity andd other
aspects off cellular phen
notype. Simplle cytotoxicity
y assays havee long been ussed, with parttial success, to predict animaal and human toxicity and to
t estimate staarting concenttrations for annimal toxicity studies.
A keey considerattion in all in vitro
v
cell-based phenotypiic assays is thhe choice of ccell type. Thee general appro
oach is to atteempt to mimiic specific hu
uman cell typ es; however, few data sugggest, for exaample,
that hepattocytes best predict
p
liver injury
i
or thatt renal tubulee cells best deetect renal injjury (Lin andd Will
2012). Asssays are cond
ducted with cells that are grown
g
in a sinngle layer (cell monolayer culture), andd these
conditionss inevitably provide
p
only a crude appro
oximation off in vivo tissuue environmennts, cell typess, and
cellcell interactions.
i
Furthermore,
F
immortalized
d (often canceer-derived) an
and other cell lines have prrovided the maainstay of celll-based assayss for decades because theyy are conveniient for obtainning large num
mbers
of cells in
n a standard state and for enabling
e
cell lines
l
to be reaadily shared bbetween labooratories. How
wever,
the cells are
a often culttured using different
d
mediia conditions,, and this cann affect cell rresponse to chemicals. For example,
e
high
h glucose con
ncentrations in
i the growthh media mightt increase thee resistance off neural cells to
o the mitocho
ondrial toxicaant 1-methyl-4
4-phenylpyriddinium (Mazzzio et al. 20100). Similarly, cytotoxicity results from assays
a
that evaluate mitochondrial dissruption can be influenceed by the Craabtree
effect, in which cancer cells preferrentially use glycolysis
g
insstead of oxiddative phosphhorylation forr ATP
production (Marroquin
n et al. 2007). Another limiitation relatedd to cell type is that many in vitro studiees use
cell lines that do not reeflect the variiation that is found
f
in the hhuman populaation. For som
me chemicalss, sensitivity to cytotoxicity varied by as much as a facctor of 100 inn a single celll type taken ffrom a broad population sam
mple (Abdo ett al. 2015).
52

There is an increasing shift away from cell lines toward cell-based assays that use cell types that are
more physiologically relevant, such as animal or human primary cells, and human induced pluripotent
stem (iPS) cells that have differentiated into specific cell types (Kraushaar et al. 2012; Godoy et al. 2013).
However, unless great care is taken to mimic tissue-relevant physical and chemical environments, those
cell types might provide little improvement over cell lines in physiological relevance. That consideration
has led to an increasing push toward the use of organotypic model systems (discussed below).
Another key technical consideration is the measurement of assay results (readout). Readouts can be
broadly divided into ones that average the response of a number of cells in a tissue culture well and ones
that assess individual cell behavior, the latter sometimes called high-content assays (discussed below in
the section Emerging Technologies). Whole-well readouts are most commonly used and have the advantage of being fast and simple, and it is straightforward to generate statistical metrics of assay performance and chemical activity. High-content readouts have the advantage, in principle, of providing much
more informationfor example, on cell-to-cell heterogeneity in response and on specific cytotoxic mechanismsbut scoring and interpreting the additional information is computationally challenging (Wink et
al. 2014).
Few studies have evaluated a large number (hundreds) of chemicals for their ability to predict acute
toxicity (OBrien et al. 2006; Xu et al. 2008; Lin and Will 2012; Porceddu et al. 2012). Most studies examine only a handful of chemicals, and studies that evaluate only specific classes of chemicals, such as endocrine disruptors or hepatotoxicants, substantially bias estimates of model performance (Thomas et al. 2012).
Cytotoxicity Assays
Measurements of cell life or death in culture have a long history in toxicology research (Ekwall
1983). Methods for measuring cytotoxicity in cell culture usually involve direct measurement of the fraction of cells that have intact membranes, for example, with neutral red uptake or fluorescent DNA dye
uptake; measurement of the metabolism of surviving cells, for example, with reduction of 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), reduction of Alomar blue, or uridine uptake; or measurement of ATP content, cell number, total DNA content, total protein content, or cell proliferation. Cytotoxicity assays are normally run for a day or more (3 days is common), so viable cells will
proliferate and increase the live-cell signal in control wells. Thus, the assays usually combine measurements of acute cell lethality, cell proliferation, or cell metabolism and might represent the simultaneous
occurrence of several mechanisms (Huang et al. 2008). The longer an assay is run, the more it tends to
factor in effects on cell proliferation unless it is run on nonproliferating cell types. Short-term (1-hour)
assays are also used but need to include sensitive measures of cell injury, such as mitochondrial membrane potential or ATP content. Some assays are based on ultraviolet (UV) or fluorescence readouts and
warrant caution because some chemicals can artificially interfere with UV- or fluorescence-based assays.
In the pharmaceutical industry, luminescent readouts have largely replaced UV- and fluorescence-based
assays.
In vitro cytotoxicity tests have been recommended as an adjunct to animal tests to improve initial
dose selection and modestly reduce the number of animals used. The registry of cytotoxicity (RC) prediction model has been recommended (NIEHS 2001a,b) for evaluating the predictive accuracy of candidate
cytotoxicity assays.4 Many RC prediction models evaluate the correlation between in vivo LD50s and in
vitro cytotoxicity IC50s.5 The BALB/c 3T3 and normal human keratinocyte (NHK) neutral red uptake
(NRU) cytotoxicity assays have been used to predict acute toxicity (NIEHS 2001a,b), but a major drawback of those cell-based assays is that they are difficult to automate. Recent advances have led to the development of commercially available ready-to-go cell plates that simply need to be defrosted and fed
with media. The plates can then be used in the conduct of high-throughput assays.
4
5
Halle (Spielmann et al. 1999; Halle 2003) describes inclusion criteria for data in the RC database.
IC50 is the chemical concentration at which 50% inhibition is achieved.
Prepublication Copy
53
Recent systematic studies that evaluated the predictiveness of in vitro assays for acute toxicity included several cell-based cytotoxicity assays partly because of their ease of use, reproducibility, and
proven (albeit limited) predictive value. For example, of the 53 assays in the European Union (EU)
ACuteTox project, seven constitute either a cell lethality assay (for example, NRU) or a metabolism assay
(for example, MTT reduction or 2-deoxyglucose and uridine uptake) that effectively measures the number
of living cells remaining after treatment. Notably, several cytotoxicity assaysan MTT assay in primary
rat hepatocytes; a cytotoxicity panel that measures intracellular Ca2+ levels, mitochondrial membrane potential, and plasma membrane potential in HepG2, SH-SY5Y and A.704 cells; and a basal cytotoxicity
NRU assay in BALB/3T3 cellsgenerated data of sufficient quality to be considered in future acutetoxicity testing strategies (Kinsner-Ovaskainen et al. 2013).
Gene-Expression and Protein-Secretion Assays
Gene-expression and protein-secretion assays aim to identify a specific toxic mechanism by measuring expression of a specific gene or secretion of a specific protein that has been implicated as a biomarker
of a particular toxic mechanism in humans. Gene-expression readouts can use engineered reporter genes,
whereby an artificial promoter drives expression of an easily assayed reporter gene (such as luciferase) in
response to activation of a toxic pathway or endogenous gene circuits. Engineered reporter genes provide
easily standardized, reliable readouts, and a panel of such assays can be used to cover multiple biological
pathways.
In recent years, several vendors have offered reporter assays for a variety of biological pathways involved in toxicity, such as inflammation, apoptosis, and endoplasmic reticulum stress. Their main limitation is that an artificial promoter in a generic cell line (such as HEK cells) might not indicate pathway
activity in the same way as the natural pathway in a specific human cell type; that is, the assays could fail
to capture biological context dependence. Reporter gene assays are of particular value for predicting endocrine disrupters because they work by reporting the expression of genes that are naturally regulated by
sex hormones. However, endocrine disruption does not qualify as a likely mechanism of systemic toxicity
that would result in an acute, debilitating injury in deployed personnel.
Expression of many endogenous genes can be simultaneously analyzed at the mRNA or protein level to create a toxicogenomic signature (discussed below under Emerging Technologies). Assays for
induction of endogenous genes can be applied to primary cells that synthesize specific proteins in response to chemicals. A major advantage of this approach over artificial promoter constructs is that it
measures endogenous gene-expression pathways and is thus more likely to indicate physiologically relevant pathway perturbation. A major disadvantage is that separate readouts must be developed for each
protein, and this makes the approach more complex and labor-intensive than measuring the expression of
a reporter like luciferase from engineered gene-expression constructs.
Protein-secretion assays are of particular value for measuring the activity of immunotoxins that induce expression and secretion of specific inflammatory cytokines, such as TNF-, IL1, and IL6 from
white blood cells. Cytokine measurements are usually made with an immune assay, typically a capture
ELISA assay that uses two antibodies to provide high specificity and sensitivity. The relevance of the cytokine assays for evaluating acute toxicity associated with chemicals of relevance to DOD is limited.
Several recent systematic studies have combined multiple high-throughput assays, including assays
for the expression of single endogenous genes in primary cells with mRNA or protein readouts. For example, the ACuteTox project included four cytokine secretion assays performed on primary white blood
cells isolated from human blood and four assays that measured synthesis of neuronal and glia proteins, at
the mRNA level, from aggregates of primary cells derived from rat brain (Kinsner-Ovaskainen et al.
2013). A meta-analysis found a variety of problems with all the assays, including failure to measure in the
correct range for immunoassays and lack of reproducibility of primary cell aggregates. The ToxCast program evaluated the performance of the Biologically Multiplexed Activity Profiling (BioMAP) human
primary cell disease models with the ToxCast phase I library (Houck et al. 2009). The chemicals that
54
Prepublication Copy

were tested generated relatively weak signatures compared with the reference pharmacological probes and
drugs, and this raised the question of whether the concentrations that were used (up to 40 M) were adequate. Furthermore, a follow-up study demonstrated that some chemicals with known biological activity,
such as pharmaceuticals, gave false-negative results in the assays (Kleinstreuer et al. 2014). All the problems highlighted here indicate the difficulty of working with complex primary cells in a high-throughput
assay format.
Limitations and Needs for Improvement of Cell-Based Phenotypic Assays
Cell-based assays are efficient in assessing chemical mechanisms of action. They have also been
useful for predicting some mechanisms of chronic toxicity, such as endocrine disruption, but might have
less applicability to acute-toxicity prediction. Furthermore, conventional cell-based cytotoxicity assays
typically lack metabolic competence and often miss chemicals that require bioactivation. They also fail to
provide data on some of the most important toxic mechanisms, notably ones that involve organ- or celltype specific physiology. For example, they miss chemicals that perturb synaptic transmission and neurotransmitter metabolism.
Responses seen in cytotoxicity assays can also be influenced by the choice of cell used. In the
Tox21 cytotoxicity profiling screen, cytotoxic response patterns varied by cell type and indicated differences in sensitivity and kinetics of the response (Xia et al. 2008). Overall, the human blood-derived cells
(Jurkat), neuron-derived cells (SH-SY5Y), and rodent cells (N2a, H-4-II-E and NIH 3T3) were most sensitive to chemical-induced cytotoxicity, and human fibroblastic, endothelial, and skin cells (HUV-EC-C,
BJ, and MRC-5) were least sensitive. Another finding from the study was the lack of similarity in the patterns of chemical activity in cells that were derived from the same tissue but from different species, for
example, human HepG2 and rat H-4-II-E hepatoma cells.
To improve the utility of cell-based assays for predicting acute toxicity, several aspects of experimental design need to be considered. First, cell-based assays that express the relevant biological processes
and toxicologic mechanisms under study need to be selected. In some cases, that might require using primary cell cultures rather than cultured immortalized cells. Second, relevant chemical concentrations and
exposure durations need to be used, and appropriate test and reference chemicals need to be identified for
evaluating the performance and predictive value of the assays. Ultimately, the readout of a cell-based assay must be mechanistically and quantitatively linked to a chemicals toxicity phenotype in humans if it is
to be truly predictive.
Organotypic Models
Many of the limitations of cell-based assays, as noted above, arise from the fact that single cells
growing on a dish are inevitably poor models for human tissues. That is the case even if the cells themselves were derived from human or animal tissue because the artificial in vitro environment changes their
phenotype in hours. The culture-induced changes affect cells responses to chemicals and the cells (especially liver cells) ability to biotransform chemicals into more or less toxic species. To address the limitation, researchers have been developing approaches for co-culture of multiple cell types or for cultures of
whole organs as slices or cell aggregates. Those approaches are collectively referred to as organotypic
models.6 The discussion below highlights a few organ systems that are relevant to acute toxicity of chemicals.
EPA defines organotypic culture models as tissue culture models that mimic in vivo tissue architecture through
interactions of heterotypic cell types (e.g., epithelium-stroma) and extracellular matrices (ECM). They can be established from isolated cells or from tissue fragments harvested in vivo, and will bridge the gap between conventional
monolayer cell cultures and whole-animal systems (EPA 2013).
Prepublication Copy
55
Skin Organotypic Models
Skin, as the largest human organ, is important for the absorption of many classical chemical-warfare
agents and constitutes the principal barrier to and defense against absorption of toxic lipophilic chemicals.
In addition, the skin is the primary site of action of acute blistering agents. The militarys historical reliance on chemical-protective boots, suits, and gloves emphasizes the importance placed on dermal protection against chemical exposure. Furthermore, one of the first questions that the warfighter asks when
told that a potent, toxic chemical can be absorbed through intact skin is, How much and how quickly?
Although definitive dermal absorption studies can be conducted on animals by using small numbers of
chemicals that are synthesized with radiolabels, various in vitro models have been developed by using
instrumented flow cells with human or porcine skin explants. Those in vitro models are far more amenable to high-throughput screening than are dermal absorption studies with radiolabeled chemicals (Basketter et al. 2012).
Organotypic cultures of skin in formats that are applicable to screening thousands of chemicals are
relatively advanced. The field has benefited from the relatively simple anatomical organization of skin,
from the fact that proliferating human keratinocytes are relatively easy to grow, from the abundant availability of human tissue from minor surgical procedures, from the need for artificial skin for treating burn
patients, and, not least, from huge investment by the cosmetics industry. Especially in the European Union, cosmetics manufacturers have been under pressure to increase safety testing while reducing animal
use. In an example that is generally encouraging for toxicity prediction, scientists at LOreal, Inc. in
France have used the EpiSkinSM model to predict irritant activity (Cotovio et al. 2005, 2008). The model
consists of primary human keratinocytes that have been induced to self-organize into a multilayered structure similar to skin by use of bioengineered substrates. MTT reduction and release of the inflammatory
cytokine IL-1 were measured. The model accurately (> 80%) predicted irritant activity of 184 cosmetic
ingredients (Cotovio et al. 2008). The study suggests that goodalthough not perfectpredictions can be
made for acute skin irritants by using a sophisticated organotypic culture model. In contrast, numerous in
vitro and in vivo models have been examined in attempts to emulate the blistering (vesication) seen in
human skin on exposure to sulfur mustard or lewisite, most with little or no success. Chemical-induced
skin blistering might be limited to humans, or it requires epithelial and fibroblast immune cell functions
that are not accounted for in current organotypic skin models.
Eye Organotypic Models
Loss of vision would be an incapacitating effect of concern to DOD. Testing for eye irritation has
benefited from investment by the cosmetics industry, although current organotypic cornea models are
generally less advanced than skin models. Commercial models are available and include the EpiOcular
OCL-200 tissues from MatTek Corporation (Ashland, MA) and the SkinEthic Reconstituted Human
Corneal Epithelium developed by a consortium of European cosmetics companies in response to banning
of rabbit testing. Systematic evaluation of the systems continues but mostly in the chemical space relevant
to cosmetics. Both systems achieved benchmarks for between-laboratory reproducibility and are undergoing tests of predictive value (Alpe et al. 2013; Pfannenbecker et al. 2013). On the basis of the results
with skin, a reasonable degree of predictive power is expected.
Lung Epithelium Organotypic Models
The lung is a relevant organ for both chemical absorption and acute toxicity. Multiple 3-D organotypic models have been described, including the commercial EpiAirway (MatTek Corporation, Ashland, MA) and MucilAir (Epithelix Srl, Geneva, Switzerland) systems. The systems use primary cells
56
Prepublication Copy

cultured at an airliquid interface, and they model airway function better than traditional submerged tissue culture. A recent paper evaluated the models relative to each other and to two conventional submerged tissue-culture systems for their ability to predict rodent lung toxicity of 19 chemicals (Sauer et al.
2015). None of the systems performed well in predicting lung toxicity in rats, and the 3-D organotypic
systems performed no better than conventional tissue-culture models. The findings suggest that organotypic lung epithelium models lag behind skin models in predictive value. The poor predictivity probably
reflects the greater complexity of airway epithelium or perhaps differences in investment.
There does not appear to be a particularly good in vitro model for lung damage associated with
chemicals (for example, phosgene) that are known to increase pulmonary permeability that results in noncardiogenic pulmonary edema. The development of in vitro assays to evaluate similar compounds identified as chemicals of interest, for example, on the basis of chemical structure or quantitative structure
activity relationships will require further DOD investment to replace animal inhalation-toxicity studies.
Liver Organotypic Models
The liver is especially relevant to biotransformation of chemicals into more or less toxic metabolites
and is an important site of acute toxic action of some chemicals. Hepatocytes are the major biochemical
engine of the liver and are responsible for metabolism and excretion of many xenobiotics. Hepatocytes
can be cultured in standard 2-D formats, but their phenotype with respect to xenobiotic metabolism and
responses changes rapidly under culture conditions. Immortalized cell lines derived from hepatocellular
carcinoma (such as HepG2) have often been used as a surrogate for hepatocytes, but their biology is even
more distant from hepatocytes in situ. Because of the relevance of liver toxicity to drug development, the
pharmaceutical industry has made major investments in modeling liver biology in 2-D cell cultures, 2-D
co-cultures, 3-D cell cultures, and engineered organotypic systems (Godoy et al. 2013).
Multiple approaches have been taken to build organotypic models of human and rodent liver, and
substantial improvements over 2-D hepatocyte culture systems in recapitulating normal liver biology,
drug metabolism, and drug responses have been noted (Khetani and Bhatia 2008; Godoy et al. 2013;
Messner et al. 2013). Despite improvements, predictive-toxicology studies still tend to focus on 2-D cultures of primary hepatocytes of hepatocellular carcinoma-derived cell lines. For example, the EU-funded
LIINTOP project is evaluating multiple 2-D culture models for liver and intestine (Gmez-Lechn et al.
2010), and the EU ACuteTox project included five assays with HepG2 cells (Kinsner-Ovaskainen et al.
2013).
The continued reliance on 2-D cultures and hepatoma-derived cell lines is problematic because testing in more robust in vitro models would probably generate more-predictive data. For example, Khetani
and Bhatia (2008) showed that a microengineered 2-D culture system in which hepatocytes are cocultured with stromal cells provided better modeling of gene expression, metabolism, and drug action
than conventional 2-D cultures. Another promising system involves co-culture of hepatocytes with
Kupffer cells in small spheroids (Messner et al. 2013). The microengineered systems often have lower
throughput and are more expensive than conventional 2-D cultures, but given the importance of the liver
in toxicology the expense might be worthwhile. Although it remains to be seen how their overall predictivity differs from that of simpler models, the use of complex liver models should improve the recognition
of inflammation and other mechanisms of toxicity that are not easily detected in simpler hepatocyte cultures. For example, Khetani et al. (2013) demonstrated that using co-cultured hepatocytes better predicted
drug-induced liver injury in a small test set of 45 chemicals.
Godoy et al. (2013) provides an exceptionally comprehensive overview of hepatic models. It is interesting to note, given that the liver is perhaps the most thoroughly characterized of the organotypic
model systems, that the authors concluded that one key message is that despite our enthusiasm for in
vitro systems, we must never lose sight of the in vivo situation. Although hepatocytes have been isolated
for decades, the hunt for relevant alternative systems has only just begun.
Prepublication Copy
57
Neural Organotypic Models
The brain has multiple neuronal and glial subtypes, complicated neuronal networks that have different types of chemical synapses, important cellcell interactions, and myelinated axons. That cellular complexity helps to make it an important site of action for many acute toxicants and makes it difficult to identify neurotoxic effects with conventional cytotoxicity assays or other in vitro systems. Brain aggregate
cultures replicate some organotypic structural and functional features and have been used as a model system for neurotoxicity testing. For example, the EU ACuteTox project included seven assays of cell aggregates derived from rat brain as part of a battery of 50 assays that included many involving 2-D cultures of
neurons (Forsby et al. 2009). The authors concluded that using aggregate cell cultures prepared from
embryonic rat brain and a multiparametric endpoint scheme, all chemicals known to be highly toxic in
humans also showed high toxicity (significant effects in the lower micromolar range) to extreme toxicity
(significant effects at nanomolar concentrations) in aggregate cultures (Honneger et al. 2009). They also
noted that inclusion of data on metabolism and pharmacokinetics of the bloodbrain barrier would likely
improve predictive value.
Limitations and Needs for Improvement of Organotypic Models
Additional organotypic models have been developed for the heart, kidney, and skeletal muscle. Organotypic models have high potential for predicting acute toxicity and potentially can recapitulate the metabolism and biological activity of a chemical. That said, the science of accurately modeling human organs in a culture dish, especially in formats suitable for high-throughput testing, is still in its infancy.
Progress is being made, but much caution is warranted, particularly for acute-toxicity prediction, which
has not been the goal of most studies. Organotypic assays are much more complex and expensive than
pure protein-based and cell-based assays and less robust than rodent models because they do not integrate
multiple physiological systems. Their reliability has often been called into question in systematic studies.
Their predictivity is also far from guaranteed; for example, the lack of success in predicting vesicant activity by using organotypic skin cultures is troubling.
For organotypic cultures to be used in DOD screening for acutely toxic chemicals, there is a need to
evaluate the potential of organotypic assays for acute-toxicity prediction and to invest further in the basic
science of organotypic cultures. The potential usefulness is high, even if current systems are far from ideal.
NONMAMMALIAN IN VIVO ANIMAL MODELS
The use of an in vivo approach facilitates the crucial understanding of how chemicals affect complex metabolic targets and pathology at the cell and organ level. Mice, rats, rabbits, and other laboratory
mammals have been used extensively to study chemical toxicity. However, in vivo assays with those species are often expensive, use large amounts of toxic test chemicals, and are difficult to use in a moderatethroughput to high-throughput manner. Those drawbacks have led toxicologists to develop alternative
animal models for chemical testing. Many alternative test organisms share biological processes with rodents and other mammals, including humans. Three test platforms of note that can be adapted to highthroughput screening rely on insect, nematode, and zebrafish models (Giacomotto and Sgalat 2010).
They can complement other cell-based in vitro test systems, and data from assays that use the alternative
animals could be used to set priorities among chemical candidates for future traditional animal testing.
Development and application of the nonmammalian models could help DOD to screen for pathwayspecific effects and could demonstrate how chemical toxicity varies among species.
This review of alternative animal models is not intended to be exhaustive; rather, the committee has
focused on end points that are relevant to acute toxicity and on selected examples that illustrate how these
systems could be adapted for high-throughput testing of acute effects.
58
Prepublication Copy

Fruit Fly Models
The fruit fly (Drosophila melanogaster) has been used as a model organism in studies of genetics and
developmental biology for over 100 years (Rubin and Lewis 2000). Fruit flies have a fully mapped genome,
and many protocols for biochemical and genetic analysis are well established. More than 60% of human
genes have functional orthologs in D. melanogaster (Bier 2005). A variety of molecular tools, including
mutagenesis and RNAi, are available for modifying fruit fly genetics. A dedicated Web-based database
(FlyBase) contains information relative to fruit fly genetics and its molecular biology (Drysdale 2008).
Fruit flies have the potential to be used for chemical-toxicity screens (Nichols 2006; Whitworth et
al. 2006; Segalat 2007). In particular, fruit flies and other insect models have improved our understanding
of the molecular action of pyrethroids, which act on both mammalian and insect sodium channels, and
other insecticides (Peterson et al. 2008). Despite their use in neurotoxicology research, there are important
limitations (Rand 2010). For example, -amino-butyric acid, acetylcholine, and other neurotransmitters
often have roles in the insect nervous system that are different from their roles in vertebrate nervous systems (Peterson et al. 2008).
Specialized video-based equipment has been developed to assess flying, chemotaxis, geotactic
climbing,7 and other behaviors (Sawin-McCormack et al. 1995; Rand 2010; Sokolowski 2001; Podratz et
al. 2011; Gregory et al. 2012; Podratz et al. 2013). In addition, eclosion8 and adult lethality are simple end
points that can be assessed without the need of a microscope. However, the transition from larva to adult
fly is complex and occurs by mechanisms distinct from those seen in mammals; this draws into question
the utility of LD50s obtained for this developmental stage (Rand 2010). There are other limitations of the
use of fruit flies. For example, chemical administration to the fly embryo must overcome the barriers presented by the hydrophobic vitelline membrane (Limbourg and Zalokar 1973; Rand 2010). In addition, fly
toxicokinetics of xenobiotics can differ substantially from that in mammals, including humans.
Nematode Models
The most widely used nematode model for biomedical research is likely Caenorhabditis elegans. C.
elegans have a fully mapped genome (C. elegans Sequencing Consortium 1998), and more than 50% of
human genes have functional orthologs in C. elegans (Harris et al. 2004). Genetic or genomic manipulationsuch as knockouts, knockdown via RNAi, and transgenic strainsis routinely available. A variety
of bioinformatics tools have been developed to support high-throughput genomic studies with this organism (Cho et al. 2014). In addition, a dedicated Web site (Wormbase) allows investigators access to microarray data and comprehensive data on gene structures, mutants, RNAi phenotypes, and proteinprotein
interactions (Chen et al. 2005).
Numerous studies have shown that C. elegans and humans share many essential biological characteristics. C. elegans has a rudimentary nervous system, exhibits behavior, and is capable of rudimentary
learning and memory functions. The anatomy of C. elegans is well understood. It contains 959 somatic
cells, including about 300 neurons that are microscopically visible. The developmental cell lineage and
the neural wiring diagram of C. elegans have been completely mapped. Many vertebrate neurotransmitters are well conserved in this nematode (Villatte et al. 1998; McVey et al. 2012). The presence of a functional nervous system has been exploited by neurotoxicologists to study the acute neurotoxic effects of
pesticides and other chemicals on this nematode (Williams and Dusenbery 1990; Ruan et al. 2009; Avila
et al. 2012; McVey et al. 2012; Meyer and Williams 2014). End points evaluated include changes in survival, behavior, locomotion, life span, cell death, neurotransmitter concentration, and function. Imagetracking systems have also been developed for assessing C. elegans locomotion (Feng et al. 2004). Melstrom and Williams (2007) reported a strong correlation between LC50s determined for C. elegans and
7
8
Drosophila instinctively climb against gravity (geotaxis).

Eclosion is the hatching of adults from the pupal stage.
Prepublication Copy
59
LD50s identified in rodents after exposure to cholinesterase-inhibiting pesticides. End points examined by
Melstrom and Williams included depression of acetylcholinesterase activity and decreased movement.
Zebrafish Models
As a vertebrate, zebrafish (Danio rerio) have substantial physiological, anatomic, and genetic homology with humans (Barbazuk et al. 2000; Howe et al. 2013; Chakravarthy et al. 2014). Zebrafish are
amenable to gene manipulation, have a short generation time, and have well-characterized rapid developmental stages; these characteristics have led to their growing use in developmental-toxicity studies (de
Esch et al. 2012; Ralda and Pia 2014). Its application to the study of developmental toxicity has garnered the interest of regulatory bodies. For example, the Organisation for Economic Co-operation and
Development (OECD) has recently formulated guidelines for using zebrafish embryos for testing acute
toxicity of 119 chemicals and for developmental toxicology (OECD 2013a,b). Zebrafish can be bred in
large numbers with minimal maintenance cost (Ralda and Pia 2014). They are a cost-effective in vivo
model for screening drugs and other chemicals and meet many objectives of a high-throughput screening
assay (Taylor et al. 2010; Tsang 2010; Lessman 2011). High-throughput zebrafish assays have also been
used for microarray and proteomic studies (Love et al. 2004). Because zebrafish larvae are transparent,
they are ideal for in vivo imaging without the use of invasive techniques (Knudsen et al. 2011; Ralda
and Pia 2014). Another advantage of the zebrafish larva model is that up to 4 days after fertilization they
are not treated as vertebrates by US institutional animal care and use committees because they retain a
yolk and higher-order neuronal functions are generally absent.
Zebrafish are used in safety pharmacology studies to screen for arrhythmogenicity (Langheinrich et
al. 2003; Milan et al. 2003; Burns et al. 2005), nephrotoxicity (Hentschel et al. 2005; Wu et al. 2012),
hepatotoxicity (Vliegenthart et al. 2014), and neurotoxicity (de Esch et al. 2012; Legradi et al. in press).
Some drugs that affect human cardiac function and structure are known to have similar effects in
zebrafish (Milan et al. 2003). Heart-specific expression of the green fluorescent protein in zebrafish has
been accomplished by using the cardiac myosin light chain 2 promoter (Huang et al. 2003). Specialized
equipment exists to monitor changes in zebrafish heart rate after chemical exposure (Burns et al. 2005;
Simoneschi et al. 2014). Behavioral assays of swimming behaviors have been developed for zebrafish
(Ali et al. 2012; Bichara et al. 2014). Driessen and co-workers (2013) have shown good concordance in
histopathological responses and gene expression profiles between zebrafish embryos and mice exposed to
known hepatotoxic chemicals. Yen et al. (2011) evaluated changes in zebrafish larva survival, acetylcholinesterase activity, and behavior after exposure to three organophosphorus pesticides. That type of study
might be useful for nerve agents and other chemicals that have a similar mechanism of action. In vivo
zebrafish assays with reverse dosimetry have also been used to develop human oral-dose hazard values
(Perkins et al. 2013).
Limitations and Needs for Improvement of Nonmammalian In Vivo Animal Models
Nonrodent animal models have the potential to assist in characterizing the acute toxicity of chemicals. One advantage of such systems is their ability to identify whole-animal and organ-level responses.
The exploration of nontraditional in vivo models for assessing acute-toxicity potential, however, will need
to consider species differences in metabolism and cellular targets and other issues related to interspecies
and in vitrotoin vivo extrapolations. Other factors to consider include differences in organ composition
(multiple cell types), cell organization or structure, and gradual enzyme expression in different tissue regions. Another challenge is related to the extrapolation of aqueous (as in the case of zebrafish) or medium
(as in the case of C. elegans) concentrations to exposure concentrations that are relevant to humans. Alternative animal model in vivo assays have considerably lower throughput than other assay systems considered by the committee and are likely to be used in the later stages of the assessment process. Little
work has been performed regarding the applicability of such models to assess the acute toxicity of chemi-
60
Prepublication Copy

cals relevant to DOD, and they have been incompletely analyzed for their predictive validity with respect
to acute toxic effects or identification of affected organ systems.
EMERGING TECHNOLOGIES
The drive to develop nonanimal methods for toxicity testing is still in its infancy, and new technologies are continually being developed. Most are aimed at commercial applications, particularly for safety
assessment in the pharmaceutical and cosmetics industries, but a subset of the new technologies will also
be useful for predicting acute toxicity. In this section, emerging technologies are broadly divided into
ones that generate large amounts of information per sample by multiplexed or image-based measurements
and ones, such as organ-on-a-chip and induced pluripotent stem (iPS) cell technologies, that aim to model
human tissues more accurately.
Multiplexed Assays
Multiplexed assays allow the measurement of dozens, hundreds, or even thousands of end points
simultaneously on a single sample. In general, they seek to provide more biological data per sample than
traditional assays that measure a single end point. Conceptually, obtaining data on many end points is expected to help in deciphering chemical mechanisms and identifying new biological targets for development of more specific assays (Larson et al. 2011). The most used, and best understood, multiplexed
readout is a gene-expression profile, in which the amounts of hundreds or thousands of mRNAs in a sample are measured in parallel (Fabian et al. 2011; Klaper et al. 2014). Microarrays have been popular for
measuring gene expression, but the decreasing cost of DNA sequencing is leading to a gradual replacement of microarray and related technologies with RNAseq approaches. There is increasing interest in
multiplexed measurement of micro-RNAs, whose expression also reflects the state of a cell or tissue.
Protein and metabolite measurements can also be multiplexed. For example, multiplexed immunoassays can be used to measure the concentration of tens or hundreds of cytokines or other proteins in a single sample. The main limitation of such assays is in developing high-quality antibody pairs to capture and
quantify a given protein with high sensitivity and specificity. Given recent developments in proteomics
technology, it is possible that mass-spectrometrybased measurements will gradually replace immunoassays for multiplexed protein measurements (Fu et al. 2010; Potts et al. 2011). Modern multiplexed proteomics methods allow quantification of thousands of proteins in tens of samples in a single spectrometry
run (McAlister et al. 2014). However, protein measurements are inevitably more difficult and less sensitive than nucleic acid measurements because proteins cannot be amplified by replication and have different physical properties and abundances. Metabolites can also be profiled by using a coupled chromatographymass spectrometry technique.
Questions remain regarding how useful multiplexed measurements will be for predicting acute toxicity, whether they are at the level of RNA, proteins, or metabolites. Most relevant information is currently available for mRNA profiling because it has the longest history, but depending on the toxic mechanism, profiling at the protein or metabolite level might be equally or more informative. Published
examples show that comparison of mRNA profiles allows identification of chemicals that have common
actions on cells in culture, including mechanisms that could cause acute toxicity (Lamb et al. 2006;
Ravindranath et al. 2015). Those examples are encouraging, but it is important to recognize that they
highlight specific success stories, not systematic toxicity prediction. Another relevant literature is on toxicogenomic approaches to predicting toxicity of chemicals in liver and other organs. Those studies typically involve treating rodents with chemicals, harvesting organs, and using microarrays combined with pathology reports to classify effects on the liver and other organs. Pharmaceutical companies and
governments have invested a great deal of resources in this approach, and major databases have recently
been released to the public. The results have been mixed: considerable improvement in understanding
toxic actions and identifying biomarkers but far from a complete solution to the problem of predicting
Prepublication Copy
61
liver toxicity (Chen et al. 2012). A recently created database includes expression signatures for 1,000
genes, using the L1000 assays, for treatment of tens of cell lines with thousands of chemicals (Duan et al.
2014). Analysis of that dataset should help to clarify the potential value of highly multiplexed geneexpression signatures in predictive toxicology.
In principle, multiplexed measurements at the protein or metabolite level might be expected to reveal a chemicals mechanisms of action more effectively, and with more predictive value, than gene-level
measurements. More ambitiously, a combination of multiplexed measurements of mRNA, protein, and
metabolite levels in parallel would in principle cover the most ground with respect to producing mechanistic information relevant to prediction of acute toxicity. That kind of integrated omics approach has
been shown to improve understanding of specific toxic mechanisms (Wilmes et al. 2013, in press) but is
expensive and unproven in its usefulness for systematic testing of acute-toxicity potential. However, the
data can be used to design new high-throughput assays that are mechanistically relevant to critical processes or pathways targeted by chemical-warfare agents.
High-Content Screening Assays
High-content screening (HCS) assays make multiple measurements of cell biology at the level of single cells by using microscopy or other imaging technologies. Typically, cells grown on multiwell plates are
treated with a chemical, stained with several fluorescent markers that report on various aspects of metabolism and organelle health, imaged, and detected with some type of automated algorithm. HCS technology
can provide much information that is relevant to specific pathways or organelles in a single assay, and it is
faster and less expensive than multiplexed assays of gene or protein expression. It has been used in drug
development for some time but only recently applied to toxicology. Recent studies show high potential
(OBrien 2014; Persson et al. 2014). In particular, some mechanisms of acute toxicity that might be poorly
detected in gene-expression assays, such as damage to cellular membranes or organelles, can be directly
assessed with HCS assays. So far, too few studies have been published to evaluate this promising technology, and key questions, such as reproducibility between laboratories, need to be addressed.
Imaging is a useful way to screen for numerous mechanistic end points at the same time, such as cell
death, apoptosis, oxidative stress, mitochondrial membrane potential, DNA damage, and cell-cycle inhibition. Algorithms have been developed, for example, to predict human liver injury. Predictivity is achieved
by testing enough positive and negative chemicals in the system to know the specificity and sensitivity of
the platform or any particular assay. In general, these assays have high specificity but low sensitivity. In
fact, in the absence of consideration of exposure, predictivity of drug-induced liver injury rarely gets
above 50% (Xu et al. 2008). In addition, because these mechanistic end points are common end points of
cell injury and death, it is difficult to predict particular organ toxicities. For example, the liver toxicant
troglitazone and the cardiac toxicant doxorubicin both cause oxidative stress and mitochondrial dysfunction in vitro. Other factors that contribute to toxicity in humanssuch as inflammation, use of multiple
drugs, and geneticsare difficult to model in simple cell systems.
Organ-on-a-Chip, Microphysiological Systems, and Advanced Organotypic Assays
The field of tissue engineering has advanced rapidly in recent years, having been stimulated especially by advances in microfabrication technology, such as micropatterning and microfluidics. The traditional goal of the tissue-engineering field is to develop replacement organs, but a shorter-term goal of
generating tissue models for drug and toxicity testing has emerged (Alepee et al. 2014; Jennings in press).
Increasingly, proponents of this kind of technology are promoting organ-on-a-chip models as the ultimate systems for determining toxicity mechanisms in organ systems (Huh et al. 2010, 2012; Esch et al.
2011; Godoy et al. 2013; NAS 2014; Pamies et al. 2014; Sung et al. 2014). Recently, Maschmeyer et al.
(in press) reported creation of long-term microphysiological systems that more closely mimic the human
liver, intestinal barrier, and skin in vivo. Those systems might have broader applications in toxicology.
62
Prepublication Copy

The considerations in evaluating the potential of these technologies are similar to those already discussed for organotypic models given that they are a modern extension of the organotypic models. Typically, organ-on-a-chip models are much more complex and expensive than simple cell-culture models. In
theory, and in some studies, their predictive value is higher than that of simple cell culture, but their reliability needs to be evaluated, and their cost per data point might exceed that of animal models, especially if
human primary cells are needed.
Induced Pluripotent Stem CellDerived Primary Cells
An iPS cell is a type of human pluripotent stem cell that can be differentiated into multiple types of
tissue cells in cell culture. iPS cells are derived from adult tissues by forced expression of stem-cell transcription factorsan approach pioneered by Shinya Yamanaka (Takahashi and Yamanaka 2006) that
avoids use of cells derived from human embryos. iPS cells can in principle provide a renewable source of
almost any human primary cell type without requiring an embryo donor. iPS cells should work well with
organotypic cultures, organ-on-a-chip technologies, and other biomimetic approaches that provide more
realistic cell-culture models (Mathur et al. 2013). Another advantage to using iPS cells, in principle, is
that they can be derived from donors who have different genotypes and might respond differently to toxicants, for example, different cytochrome p450 alleles that cause differences in drug metabolism. Thus, the
long-term potential for application of iPS cells to toxicity testing is high. Nevertheless, many hurdles
must be overcome, including development of methods for reliable differentiation of iPS cells into cell
types that are relevant to acute toxicity. The field is promising but unlikely to be useful for chemical testing in the next 5 years. One cell type that is relevant to acute toxicity and is relatively easy to generate
from iPS cells is human cardiomyocytes (Kraushaar et al. 2012); cardiotoxicity testing will therefore be a
field to monitor for application of iPS technology to toxicology.
METABOLIC CONSIDERATIONS
Chemical metabolism is an important in vivo biological process that should be considered during the
interpretation of in vitro testing data. There is a large capacity for metabolism in the body; metabolizing
enzymes are present in the liver, lung, nasal region, and other tissues. Metabolism can convert parent
chemicals into toxic metabolites (metabolic activation), into nontoxic metabolites (metabolic inactivation), or into metabolites that are rapidly removed from circulation (detoxification). Some in vitro systems
are metabolically competent and thus provide at least some metabolism in an assay. Others are less metabolically competent and should use parent chemicals and metabolites as test agents to ensure that the appropriate chemicals are assessed. The use of structural identification tools described in Chapter 3 can help
to predict toxic metabolites. However, it can be difficult to identify a biologically active molecule without
knowing the metabolic pathway in advance. This section summarizes metabolic assays and considerations
that can complement in vitro testing strategies that focus on activity of a parent chemical. Although the
most effective classical chemical-warfare agents do not require metabolic activation, a clear understanding of the role of chemical metabolism in toxicity has the potential to refine acute-toxicity assessments
when incorporated with other testing strategies.
Assessing Reactive-Metabolite Formation
Many structural alerts9 that indicate the formation of reactive metabolites have been identified. Reactive metabolite formation, however, is not necessarily sufficient for toxicity to occur; in fact, many
marketed drugs, which are considered safe at therapeutic dosages, contain such alerts (Kalgutkar and
As noted in Chapter 3, a structural alert is a chemical structure that has been linked to toxicity or a specific toxicity end point.
Prepublication Copy
63
Dalvie 2014). Thus, it is difficult to assign toxicity to metabolites in the absence of in vitro experimentation. An initial strategy that could be used in the assessment of metabolite toxicity could be to identify
potential metabolites by nontesting approaches (such as the use of quantitative structureactivity relationships) and then to confirm the presence of predicted chemical moieties experimentally. Formation of human metabolites can be assessed with in vitro systems (such as incubations with liver microsomes or
hepatocytes) coupled with mass-spectrometric detection and measurement of metabolite formation. Thus,
screening for toxicity can be based directly on a known metabolite or determined analytically by measuring metabolites formed during an assay.
Some chemical moieties (such as quinones and epoxides) are known to elicit toxicity. Highly electrophilic metabolites are known to be reactive with reduced glutathione (GSH), an endogenous nucleophile that plays a key role in xenobiotic metabolism and detoxification. Detection of GSH conjugates in in
vitro assays that incorporate hepatocytes or liver microsomes provide indirect evidence that a reactive
metabolite was formed.
In Vitro Systems to Study the Role of Metabolism in Toxicity Potential
A variety of in vitro model systems have been developed to study metabolism and include precision-cut tissue slices, subcellular fractions such as the microsomal fraction, primary cells in suspension,
primary monolayers of cells in culture, continuous cell lines, immortalized primary cells, liver-derived
cell lines re-expressing biotransformation enzymes and genetically-engineered cell lines expressing biotransformation enzymes (Combes et al. 2006). Tissue fractions that contain metabolic enzymes, such as
microsomes or S9 fractions, can also be introduced into an assay system to increase its metabolic competence (Glatt et al. 1989). Encapsulation of S9 in hydrogel microbeads has been recently introduced into
cytotoxicity assays as one method of reducing leakage of potentially toxic microsomal lipid peroxides
(Yamamoto et al. 2011).
Despite the availability of in vivo rodent acute-toxicity data on some chemicals, the studies will not
identify toxic metabolites and resulting toxicity that are elicited via human-specific metabolic pathways
not present in laboratory animal species. One approach to evaluating species differences in metabolism is
to use primary human hepatocytes or other human-origin in vitro systems. Another approach to elucidating metabolic pathways relies on evaluating cell responses in the presence and absence of a P450 inhibitor, such as 1-aminobenzotriazole (ABT). If the formation of metabolites is inhibited by ABT, isoformspecific inhibitors can be used to identify specific isoforms involved. An individual cDNA-expressed
human P450 isoforms system can be used to confirm the results of P450 isoform-specific inhibitors.
Current State of Metabolic Competence in in Vitro Testing Approaches
Because of their robust proliferative potential and sensitivity to toxic effects, immortalized cell lines
have been identified as the system of choice for high-throughput assays. However, most of the cell lines
have little or no metabolic enzyme content (that is, they are metabolically incompetent) and respond differently from tissue slices, primary cells, or tissue that is exposed to the same stimuli in vivo. For example, breast-cancer cell lines, such as MCF-7, are more prone to toxic effects than normal breast cells. Furthermore, although the hepatoma-derived HepG2 cells have many liver-specific functions and express
conjugating enzymes, they lack functional expression of almost all the relevant human xenobiotic metabolizing enzymes in the cytochrome P450 family (Donato et al. 2008). Primary cells that are derived
directly from animal or human tissues are more metabolically competent than immortalized cell lines but
have much shorter half-lives and require much more care to be sustained for toxicity screening. More recent advances in cell-culture technology have improved metabolic capability in in vitro systems. HepaRG
cells cultured in 3-D spinner-bioreactors are an attractive tool for toxicological studies and show an
64
Prepublication Copy

expression of CYP450 enzymes and phase II metabolism that more closely mimics in vivo conditions
(Leite et al. 2012). Multicellular 3-D human primary liver cell cultures that contain hepatocytes, fibroblasts, stellate cells, and Kupffer cells have also demonstrated increased metabolic activity in the presence
of fluid flow (Esch et al. 2015). In summary, cell systems have their own advantages and drawbacks, and
understanding their enzymatic content will be important for estimating the effect of metabolism on in
vitro screening data.
In the absence of metabolic activity, in vitro assays might still be used effectively in screening for
biological activity of the parent chemical. However, in vitro assays designed to assess chemical metabolism or enzyme involvement might not translate directly to in vivo effects in that they might not recapitulate the endogenous concentrations or physiological distribution of enzymes in vivo. Although highthroughput screening focuses on targeted assay systems, pharmacokinetic information can be integrated
independently, as further discussed in Chapter 5.
ASSAY CONSIDERATIONS FOR IMPROVING PREDICTION OF ACUTE TOXICITY
The limitations of specific in vitro assays have been discussed above, and here the focus is on broad
steps that could be taken to improve the ability of screening assays to predict acute toxicity. The largest
improvement needed is the demonstration of a linkage of assay measurements to relevant mechanisms of
toxicity that quantitatively reflect an in vivo toxicity phenotype in target cell types. Prediction of acute
toxicity would also be improved if the route of exposure were considered in assay design. The committee
assumes that the relevant exposure routes are dermal and inhalation, and most assays have been designed
to model oral and intravenous exposures. The issues surrounding exposure route increase as assays
become higher throughput and less metabolically competent. As mentioned earlier, future assays should
also take chemical metabolism into account. Those and other considerations are discussed in more detail
below.
Quantitative Linkage of Assay Measurements to in Vivo Phenotype
Validated alternative test methods are needed for evaluating the safety of chemicals, cosmetics, and
drugs. To address that need, the EU ACuteToX program assessed the ability of in vitro and in silico tools
and assays to predict specific organ and system toxicity (such as hematotoxicity, neurotoxicity, nephrotoxicity, and hepatotoxicity) and intestinal absorption, distribution, and metabolism. The first prevalidation phase of the ACuteTox project tested a set of 57 chemicals in 50 in vitro and in silico assays or approaches and used published toxicity data as the phenotypic anchor for statistical analyses. That phase
identified eight assays that showed some value in predicting acute toxicity (Clemedson et al. 2006); Box
4-1 provides more detailed information. The ACuteTox project identified several broad kinds of improvement that would need to be considered in the design of future in vitro screening systems, namely,
improved consideration of mechanistic data and increased use of pharmacokinetic data to enhance toxicity estimates (ACuteTox 2010).
One example of mechanistically informed assay design comes from the field of testing of mitochondrial toxicity. Mitochondrial toxicity is a major contributor to organ toxicity, such as toxicity in liver, kidney, heart, muscle, and the central nervous system (Dykins and Will 2008). Acute effects can be measured rapidly in 96-well formats by using isolated mitochondrial and soluble-oxygen sensor technology
(Luxcel Biosciences, Cork, Ireland) or cell-based assays that evaluate mitochondrial respiration and glycolysis (Seahorse Biosciences, Bellerica, MA) (Will and Dykens 2014). Those assays potentially can be
multiplexed with readouts of mitochondrial membrane potential dissipation or cytotoxicity (ATP content), and incubation times can be tailored to be from 124 hours after drug exposure (Porceddu et al.
2012). In the absence of pharmacokinetic data, most biochemical and cell-based assays remain primarily
ranking tools for hazard identification, not predictors of true risk.
Prepublication Copy
65
BOX 4-1 ACuteToX Testing Strategy

The goal of the EU ACuteTox project was to evaluate whether regulatory animal tests for acute systemic toxicity could be replaced with a combination of in vitro assays. The ACuteToX program assessed the correlation of concentrations for in vitro activity with effective doses derived from wholeanimal studies and evaluated a series of assays and physicochemical properties to determine how
well they predicted in vivo acute systemic toxicity. On the basis of statistical analysis, eight test methods were found to be promising for inclusion in the testing strategy and, therefore, selected for participation in the pre-validation study:
The neutral red uptake assay that uses the 3T3 fibroblast cell line (3T3/NRU).
The cytokine release assay that uses human whole blood (IL-1, IL-6, and TNF-).
Inhibition of colony-forming-unit efficiency in human cord bloodderived cells stimulated with
CFU-GM (CBC/CFU-GM).
Gene expression (GFAP, HSP-32, MBP, and NF-H) and uridine incorporation measuring total
mRNA synthesis in primary rat brain aggregate cultures.
A panel that measures oxidative stress (intracellular peroxidative activity, intracellular concentrations of superoxide anion, and oxidized DNA base 8-oxoguanine) and cytotoxicity screening (intracellular Ca2+ concentrations, mitochondrial membrane potential, and plasma membrane potential) in HepG2, SH-SY5Y, and A.704 cells.
The MTT assay that uses primary rat hepatocytes.
Kinetic parameters (volume of distribution, protein binding, clearance, and oral absorption using
Caco-2 cells and neuronal networks) for estimating the oral dose on the basis of the effective
concentration observed in vitro.
Estimation of chemical passage through the bloodbrain barrier using neuronal networks (for
neurotoxicity assays).
Sources: ACuteTox (2010); Combes et al. (2006).
Some in vitro assays have been sufficiently validated that they can largely replace animal testing.
For example, the European Centre for the Validation of Alternative Methods (ECVAM) Scientific Advisory Committee endorsed the EPISKIN test and the EpiDerm method as scientifically valid replacements
in a tiered testing strategy for the rabbit skin irritation method and for identifying skin irritants, respectively (Spielmann et al. 2007). In addition, the use of the Cultex Radial Flow System to assess acute
pulmonary toxicity of fine dusts and nanoparticles could possibly be adapted to test for chemicals that
might be relevant to DOD (Steinritz et al. 2013). In vivoin vitro comparison of acute respiratory tract
toxicity using human 3-D airway epithelial models and human A549 and murine 3T3 monolayer cell systems has also been reported (Sauer et al. 2013).
In summary, evaluations of in vitro assays for predicting acute toxicity have focused on nonmechanistic indicators of toxicity, such as cytotoxicity assays, or low-throughput measurements. Few in vitro
assays have been developed with quantitative linkage to any phenotype (acute or chronic). Screening for
acute toxicity by using mechanisms that are likely to cause debilitating injury (Table 2-1) will require assays that are purposefully selected for their biochemical targets and characterized for their potential value
in predicting human toxicity.
Assay Considerations: Lessons Learned from High-Throughput Screening Programs
The recent investment in EPAs ToxCast program and the broader Tox21 Initiative has allowed important progress in the development and use of HTS platforms to assess biological activity and potential
mechanisms of action for industrial and environmental chemicals. Such programs provide rapid screening
66
Prepublication Copy

of hundreds of chemicals for dozens of cellular targets and relevant pathways. However, some chemical,
cellular, and assay conditions need to be considered if one is to use and interpret the data appropriately.
First, for various reasons, the nominal chemical concentration added to the assay well is not necessarily representative of the concentration at which chemical bioactivity is observed (Groothuis et al. in press).
Chemical purity must be confirmed and solubility in the assay medium checked to determine that the initial
applied concentration is accurately known. Chemical stability also needs to be monitored over the course of
the assay so that chemical stability and availability can be tracked. Labile chemicals can degrade rapidly on
exposure to light, aqueous conditions, or other constituents of the media or in vitro system. In fact, MacArthur et al. (2009) conducted chemical stability studies with cytochrome P450 assays at NCGC and found
decreased chemical potency over time and lower efficacy of older samples stored in dimethyl sulfoxide
(DMSO). For that reason, test chemicals at NCGC are used for no longer than 46 months. If degradation
does occur, assay bioactivity (or lack thereof) might be inaccurately attributed to the parent chemical. Alternatively, the chemical might bind to plastics, cellular constituents, or proteins in the in vitro system and render it unavailable to elicit any effect in the test system (Blaauboer 2010).
Second, there are limitations of current cellular systems. The cells used in HTS assays typically are
selected for their proliferative capacity, adherent properties, and ease of growth in high-throughput plates
and systems (Shukla et al. 2010). Immortalized cancer cell linessuch as MCF-7 (breast cancer), A549
(lung cancer), and HepG2 (liver cancer)are commonly used. It is possibleor perhaps likelythat assessments in the limited cellular space might fail to detect chemical activities or effects that might occur
in normal (nontumor) differentiated cells. In addition, proliferative cell lines have a reduced ability to metabolize parent chemicals. To address those issues, new hepatic cell lines are being developed to be more
metabolically competent and are discussed further in the next paragraph..
Third, assay reproducibility can be an issue. Chemical autofluorescence and cytotoxicity are common
causes of assay interference that can lead to false positives or false negatives (Huang et al. 2011). Furthermore, cells can have different levels of activity or responsiveness, depending on whether they are primary
cells, differentiated cells, or immortalized cells and on how many times they have been passaged.10 Variability in metabolic capabilities among various sources of isolated primary hepatocytes is well documented and
is due to numerous factors, including isolation issues and donor variability. Recently, the HepaRG cell line
has been introduced as a hepatic cell line that has a degree of metabolic competence and is amenable to use
in a 96-well high-throughput system (Guillouzo et al. 2007). However, maintenance of the HepaRG cells in
a differentiated state requires the use of high concentrations of DMSO, which has substantial cellular effects, including inhibition of metabolizing enzymes of the cytochrome P450 family and alteration of membrane permeability and antioxidant status. Whereas the HepaRG cells are recognized as metabolically competent, the competence is on a much lower scale than that of fresh or cryopreserved primary hepatocytes
(Kanebratt and Andersson 2008; Lubberstedt et al. 2011).
Fourth, interpretation of activity or effective concentrations from HTS assays should be carefully
considered. The concentration at which bioactivity is observed should be considered in the context of the
complete activity profile among all assays tested and the range of the dose-response relationships. Review
of the ToxCast data has revealed that a burst of activity in many assays might result at concentrations
close to or approaching cytotoxicity (EPA 2014b). Activity measurements at high concentrations probably represent nonspecific effects and offer little information about specific bioactivity. Likewise, a lack of
response can be due to tested concentrations below bioactivity, lack of representation of the biological
target, or assay unreliability. The finding that cytotoxicity of some chemicals varies with the cell type
emphasizes the need to characterize biological activity over a broad concentration range (Xia et al. 2008).
Fifth, assays should always include positive-control reference chemicals, whose activity can be used
to determine assay variability, sensitivity, and specificity. Some of those considerations were discussed
by Judson et al. (2013) and Patlewicz et al. (2013). Indeed, efforts to characterize and document
10
Cultured cells are routinely transferred and replated (subcultured) to avoid the senescence associated with high
cell density. Passage number refers to the number of times that the cells have been subcultured.
Prepublication Copy
67
nonguideline in vitro assays, including high-throughput and high-content assays, have been made under
the auspices of OECD, which has recently published a guidance document (OECD 2014).
Finding: On the basis of the results of HTS programs, in vitro assays have demonstrated some
predictive value for acute toxicity; therefore, an in vitro screening approach for predicting the potential
for chemicals to cause acute, debilitating injuries is theoretically feasible.
Finding: Current assays were not developed for predictions of acute toxicity (particularly lethality) and have generally not dealt with chemicals that are acutely toxic or debilitating. The few evaluations
of in vitro assays to predict acute toxicity have focused primarily on nonmechanistic indicators of toxicity, such as cytotoxicity assays, or on low-throughput measurements.
Finding: There is little evidence that results of in vitro assays are predictive of in vivo outcomes
of concern to DOD. A screening program for acute toxicity will require the development of new in vitro
assays that are mechanistically relevant to critical processes or pathways that are related to acute, debilitating toxicity.
Finding: Evaluations of in vitro assays have focused on oral and intravenous exposure. The evaluation and development of in vitro assays that address dermal and inhalation exposures and contact toxicity will require additional research to understand absorption and permeability in the skin and lung.
Finding: Most in vitro assays do not account for important pharmacokinetic characteristics, such
as metabolism, that can influence in vivo toxicity. Although in vitro assays lacking metabolic capacity
can effectively screen for biological activity of the parent chemical, the pharmacokinetic relationship between exposure and concentration at a target site needs to be addressed.
Finding: In vitro testing and screening programs (Tox21, ToxCast, and ACuteTox) offer a number of useful lessons regarding assay reliability, chemical solubility or purity, standardized reference
chemicals, doseresponse experimental designs, and standardized data processing.
Recommendation: The experience of HTS programs should be considered in the design of an in
vitro screening program to predict acute, debilitating toxicity. Because of the potential need to include
highly toxic agents, if only as reference chemicals, such a screening program will need to consider health,
safety, and environmental issues associated with handling highly toxic and threat agents.
REFERENCES
Abdo, N., M. Xia, C.C. Brown, O. Kosyk, R. Huang, S. Sakamuru, Y.H. Zhou, J. Jack, P. Gallins, K. Xia, Y. Li,
W.A. Chiu, A. Motsinger-Reif, C.P. Austin, R.R. Tice, I. Rusyn, and F.A. Wright. 2015. Population-based in
vitro hazard and concentration-response assessment of chemicals: The 1000 Genomes High-Throughput
Screening Study. Environ. Health Perspect. 125(5):458-466.
ACuteTox. 2010. ACuteTox: Prevalidation, Highlights and Main Achievements of the Project. News Letter [online].
Available: http://www.acutetox.eu/Final_newsletter_Oct_2010.pdf [accessed May 7, 2015].
Alpe, N., S. Bessou-Touya, J. Cotovio, A. de Smedt, B. de Wever, C. Faller, P. Jones, B. Le Varlet, M. MarrecFairley, U. Pfannenbecker, M. Tailhardat, F. van Goethem, and P. McNamee. 2013. Cosmetics Europe multilaboratory pre-validation of the SkinEthic reconstituted human corneal epithelium test method for the prediction of eye irritation. Toxicol. In Vitro 27(5):1476-1488.
Alepee, N., A. Bahinski, M. Daneshian, B. De Wever, E. Fritsche, A. Goldberg, J. Hansmann, T. Hartung, J. Haycock, H. Hogberg, L. Hoelting, J.M. Kelm, S. Kadereit, E. McVey, R. Landsiedel, M. Leist, M. Lubberstedt,
F. Noor, C. Pellevoisin, D. Petersohn, U. Pfannenbecker, K. Reisinger, T. Ramirez, B. Rothen-Rutishauser, M.
Schafer-Korting, K. Zeilinger, and M.G. Zurich. 2014. State-of-the-art of 3D cultures (organs-on-a-chip) in
safety testing and pathophysiology. ALTEX 31(4):441-477.
Ali, S., D.L. Champagne, and M.K. Richardson. 2012. Behavioral profiling of zebrafish embryos exposed to a panel
of 60 water-soluble compounds. Behav. Brain Res. 228(2):272-283.
68
Prepublication Copy

Attene-Ramos, M.S., N. Miller, R. Huang, S. Michael, M. Itkin, R.J. Kavlock, C.P. Austin, P. Shinn, A. Simeonov,
R.R. Tice, and M. Xia. 2013. The Tox21 robotic platform for the assessment of environmental chemicals from vision to reality. Drug Discov. Today. 18(15-16):716-723.
Austin, C., R. Kavlock., and R. Tice. 2008. Tox21: Putting a Lens on the Vision of Toxicity Testing in the 21st Century [online]. Available: http://alttox.org/tox21-putting-a-lens-on-the-vision-of-toxicity-testing-in-the-21st-cen
tury/ [accessed March 19, 2014].
Avila, D., K. Helmcke, and M. Aschner. 2012. The Caenorhabiditis elegans model as a reliable tool in neurotoxicology. Hum. Exp. Toxicol. 1(3):236-243.
Barbazuk, W.B., I. Korf, C. Kadavi, J. Heyen, S. Tate, E. Wun, J.A. Bedell, J.D. McPherson, and S.L. Johnson.
2000. The syntenic relationship of the zebrafish and human genomes. Genome Res. 10(9):1351-1358.
Basketter, D., D. Jrova, and H. Kandrov. 2012. Review of skin irritation/corrosion hazards on the basis of human
data: A regulatory perspective. Interdiscip. Toxicol. 5(2):98-104.
Bichara, D., N.B. Calcaterra, S. Arranz, P. Armas, and S.H. Simonetta. 2014. Set-up of an infrared fast behavioral
assay using zebrafish (Danio rerio) larvae, and its application in compound biotoxicity screening. J. Appl.
Toxicol. 4(2):214-219.
Bier, E. 2005. Drosophila, the golden bug, emerges as a tool for human genetics. Nat. Rev. Genet. 6(1):9-23.
Blaauboer, B.J. 2010. Biokinetic modeling and in vitro-in vivo extrapolations. Toxicol. Environ. Health B Crit. Rev.
13(2-4):242-252.Burns, C.G., D.J. Milan, E.J. Grande, W. Rottbauer, C.A. MacRae, and M.C. Fishman. 2005.
High-throughput assay for small molecules that modulate zebrafish embryonic heart rate. Nat. Chem. Biol.
1(5):263-264.
Burns, C.G., D.J. Milan, E.J. Grande, W. Rottbauer, C.A. MacRae, and M.C. Fishman. 2005. High-throughput assay
for small molecules that modulate zebrafish embryonic heart rate. Nat. Chem. Biol. 1(5):263-264.
C. elegans Sequencing Consortium. 1998. Genome sequence of the nematode C. elegans: A platform for investigating biology. Science 282(5396):2012-2018.
Chakravarthy, S., S. Sadagopan, A. Nair, and S.K. Sukumaran. 2014. Zebrafish as an in vivo high-throughput model
for genotoxicity. Zebrafish 11(2):154-166.
Chen, M., M. Zhang, J. Borlak, and W. Tong. 2012. A decade of toxicogenomic research and its contribution to toxicological science. Toxicol. Sci. 130(2):217-228.
Chen, N., T.W. Harris, I. Antoshechkin, C. Bastiani, T. Bieri, D. Blasiar, K. Bradnam, P. Canaran, J. Chan, C.K.
Chen, W.J. Chen, F. Cunningham, P. Davis, E. Kenny, R. Kishore, D. Lawson, R. Lee, H.M. Muller, C.
Nakamura, S. Pai, P. Ozersky, A. Petcherski, A. Rogers, A. Sabo, E.M. Schwarz, K. Van Auken, Q. Wang, R.
Durbin, J. Spieth, P.W. Sternberg, and L.D. Stein. 2005. WormBase: A comprehensive data resource for Caenorhabditis biology and genomics. Nucleic Acids Res. 33(Database Issue):D383-D389.
Cho, A., J. Shin, S. Hwang, C. Kim, H. Shim, H. Kim, H. Kim, and I. Lee. 2014. WormNet v3: A network-assisted
hypothesis-generating server for Caenorhabditis elegans. Nucleic Acids Res. 42(Web Server issue):W76-W82.
Clemedson, C. 2008. The European ACuteTox project: A modern integrative in vitro approach to better prediction
of acute toxicity. Clin. Pharmacol. Ther. 84(2):200-202.
Clemedson, C., B. Blaauboer, J.V. Castell, P. Prieto, L. Risteli, J.A, Vericat, and A. Wendel. 2006. ACuteTox
Optimation and pre-validation of an in-vitro test strategy for predicting human acute toxicity. ALTEX
23(Suppl.):254-258.
Clemedson, C., A. Kolman, and A. Forsby. 2007. The integrated acute systemic toxicity project (ACuteTox) for the
optimisation and validation of alternative in vitro tests. ATLA 35(1):33-38.
Combes, R., S. Coecke, M. Jacobs, E. Brandon, U. Bernauer, and W. Janssens. 2006. The Use of Metabolizing Systems for in Vitro Testing of Endocrine Disruptors [online]. Available: http://www.oecd.org/chemicalsafety/
testing/36280446.pdf [accessed May 7, 2015].
Cotovio, J., M.H. Grandidier, P. Portes, R. Roguet, and G. Rubinstenn. 2005. The in vitro skin irritation of chemicals: Optimisation of the EPISKIN prediction model within the framework of the ECVAM validation process.
ATLA 33(4):329-349.
Cotovio, J., M. Grandidier, D. Lelivre, R. Roguet, E. Tinois-Tessonneaud, and J. Leclaire. 2008. In vitro acute skin
irritancy of chemicals using the validated EPISKIN model in a tiered strategy: Results and performances with
184 cosmetic ingredients. AATEX 14(Special Issue):351-358. [online]. Available: http://altweb.jhsph.edu/
wc6/paper351.pdf [accessed March 19, 2015].
de Esch, C., R. Slieker, A. Wolterbeek, R. Woutersen, and D. de Groot. 2012. Zebrafish as potential model for
developmental neurotoxicity testing: A mini review. Neurotoxicol. Teratol. 34(6):545-553.
Dix, D.J., K.A. Houck, M.T. Martin, A.M. Richard, R.W. Setzer, and R.J. Kavlock. 2007. The ToxCast program for
prioritizing toxicity testing of environmental chemicals. Toxicol. Sci. 95(1):5-12.
Prepublication Copy
69
Donato, M.T, A. Lahoz, J.V. Castell, and M.J. Gomez-Lechon. 2008. Cell lines: A tool for in vitro drug metabolism
studies. Curr. Drug Metab. 9(1):1-11.
Driessen, M., A.S. Kienhuis, J.L. Pennings, T.E. Pronk, E.J. van de Brandhof, M. Roodbergen, H.P. Spaink, B. van
de Water, and L.T. van der Ven. 2013. Exploring the zebrafish embryo as an alternative model for the evaluation of liver toxicity by histopathology and expression profiling. Arch. Toxicol. 87(5):807-823.
Drysdale, R. 2008. FlyBase: A database for the Drosophila research community. Methods Mol. Biol. 420:45-59.
Duan, Q., C. Flynn, M. Niepel, M. Hafner, J.L. Muhlich, N.F. Fernandez, A.D. Rouillard, C.M. Tan, E.Y. Chen,
T.R. Golub, P.K. Sorger, A. Subramanian, and A. Maayan. 2014. LINCS Canvas Browser: Interactive web
app to query, browse and interrogate LINCS L1000 gene expression signatures. Nucleic Acids Res. 42(Web
Server Issue):W449-W460.
Dykens, J.A., and Y. Will. 2008. Drug-Induced Mitochondrial Dysfunction. Hoboken, NJ: Wiley.
Ekwall, B. 1983. Screening of toxic compounds in mammalian cell cultures. Ann. N.Y. Acad. Sci. 407:64-77.
EPA (US Environmental Protection Agency). 2003. Framework for Computational Toxicology Research Program in
ORD. EPA/600/R-03/065. Office of Research and Development, US Environmental Protection Agency,
Washington, DC [online]. Available: http://epa.gov/comptox/download_files/basic_information/comptoxfra
mework06_02_04.pdf [accessed May 6, 2015].
EPA (US Environmental Protection Agency). 2013. Organotypic Culture Models for Predictive Toxicology Center:
Synopsis of Program [online]. Available: http://epa.gov/ncer/rfa/2013/2013_star_ocm.html [accessed March
19, 2015].
EPA (US Environmental Protection Agency). 2014a. ToxCast Assay Annotation Version 1.0 Data User Guide. National Center for Computational Toxicology, US Environmental Protection Agency, Research Triangle Park,
NC. October 2014 [online]. Available: http://epa.gov/ncct/toxcast/files/ToxCast%20Assays/ToxCast_Assay
_Annotation_Data_Users_Guide_20141021.pdf [accessed March 19, 2015].
EPA (US Environmental Protection Agency). 2014b. Meeting on Scientific Issues Associated with Integrated Endocrine Bioactivity and Exposure based Prioritization and Screening, December 2-5, 2014. FIFRA Scientific
Advisory Panel, US Environmental Protection Agency [online]. Available: http://www.epa.gov/scipoly/sap/
meetings/2014/december/120214transcript.pdf [accessed March 26, 2015].
Esch, M.B., T.L. King, and M.L. Shuler. 2011. The role of body-on-a-chip devices in drug and toxicity studies. Annu. Rev. Biomed. Eng. 13:55-72.
Esch, M.B., J.M. Prot, Y.I. Wang, P. Miller, J.R. Llamas-Vidales, B.A. Naughton, D.R. Applegate, M.L. Shuler.
2015. Multi-cellular 3D human primary liver cell culture elevates metabolic activity under fluidic flow. Lab
Chip. 15(10):2269-2277.
Fabian, G., N. Farago, L.Z. Feher, L.I. Nagy, S. Kulin, K. Kitajka, T. Bito, V. Tubak, R.L. Katona, L. Tiszlavicz,
and L.G. Puskas. 2011. High-density real-time PCR-based in vivo toxicogenomic screen to predict organspecific toxicity. Int. J. Mol. Sci. 12(9):6116-6134.
Feng, Z., C.J. Cronin, J.H. Wittig, Jr., P.W. Sternberg, and W.R. Schafer. 2004. An imaging system for standardized
quantitative analysis of C. elegans behavior. BMC Bioinformatics 5:115.
Forsby, A., A.K. Bal-Price, A. Camins, S. Coecke, N. Fabre, H. Gustafsson, P. Honegger, A. Kinsner-Ovaskainen,
M. Pallas, V. Rimbau, E. Rodrguez-Farr, C. Suol, J.A. Vericat, and M.G. Zurich. 2009. Neuronal in vitro
models for the estimation of acute systemic toxicity. Toxicol. In Vitro. 23(8):1564-1569.
Fu, Q., F.S. Schoenhoff, W.J. Savage, P. Zhang, and J.E. Van Eyk. 2010. Multiplex assays for biomarker research
and clinical application: Translational science coming of age. Proteomics Clin. Appl. 4(3):271-284.
Giacomotto., J, and L. Sgalat. 2010. High-throughput screening and small animal models, where are we? Br. J.
Pharmacol. 160(2):204-216.
Glatt, H., R. Padykula, G.A. Berchtold, G. Ludewig, K.L. Platt, J. Klein, and Oesch. 1989. Multiple activation pathways of benzene leading to products with varying genotoxic characteristics. Environ. Health Perspect. 82:81-89.
Godoy, P., N.J. Hewitt, U. Albrecht, M.E. Andersen, N. Ansari, S. Bhattacharya, J.G. Bode, J. Bolleyn, C. Borner, J.
Bttger, A. Braeuning, R.A. Budinsky, B. Burkhardt, N.R. Cameron, G. Camussi, C.S. Cho, Y.J. Choi, J.C.
Rowlands, U. Dahmen, G. Damm, O. Dirsch, M.T. Donato, J. Dong, S. Dooley, D. Drasdo, R. Eakins, K.S.
Ferreira, V. Fonsato, J. Fraczek, R. Gebhardt, A. Gibson, M. Glanemann, C.E.P. Goldring, M.J. GmezLechn, G.M.M. Groothuis, L. Gustavsson, C. Guyot, D. Hallifax, S. Hammad, A. Hayward, D. Hussinger,
C. Hellerbrand, P. Hewitt, S. Hoehme, H.G. Holzhtter, J.B. Houston, J. Hrach, K. Ito, H. Jaeschke, V. Keitel,
J.M. Kelm, B.K. Park, C. Kordes, G.A. Kullak-Ublick, E.L. LeCluyse, P. Lu, J. Luebke-Wheeler, A. Lutz,
D.J. Maltman, M. Matz-Soja, P. McMullen, I. Merfort, S. Messner, C. Meyer, J. Mwinyi, D.J. Naisbitt, A.K.
Nussler, P. Olinga, F. Pampaloni, J. Pi, L. Pluta, S.A. Przyborski, A. Ramachandran, V. Rogiers, C. Rowe, C.
Schelcher, K. Schmich, M. Schwarz, B. Singh, E.H.K. Stelzer, B. Stieger, R. Stber, Y. Sugiyama, C. Tetta,
70
Prepublication Copy

W.E. Thasler, T. Vanhaecke, M. Vinken, T.S. Weiss, A. Widera, C.G. Woods, J. J. Xu, K.M. Yarborough, and
J.G. Hengstler. 2013. Recent advances in 2D and 3D in vitro systems using primary hepatocytes, alternative
hepatocyte sources and non-parenchymal liver cells and their use in investigating mechanisms of hepatotoxicity, cell signaling and ADME. Arch. Toxicol. 87(8):1315-1530.
Gmez-Lechn, M.J., L. Tolosa, J.V. Castell, and M.T. Donato. 2010. Mechanism-based selection of compounds for
the development of innovative in vitro approaches to hepatotoxicity studies in the IIINTOP project. Toxicol.
In Vitro 24(7):1879-1889.
Gregory, J.M., T.P. Barros, S. Meehan, C.M. Dobson, and L.M. Luheshi. 2012. The aggregation and neurotoxicity
of TDP-43 and its ALS-associated 25 kDa fragment are differentially affected by molecular chaperones in
Drosophila. PLoS One 7(2):e31899.
Groothuis, F.A., M.B. Heringa, B. Nicol, J.L. Hermens, B.J. Blaauboer, and N.I. Kramer. In press. Dose metric considerations in in vitro assays to improve quantitative in vitro-in vivo dose extrapolations. Toxicology.
Guillouzo, A., A. Corlu, C. Aninat, D. Glaise, F. Morel, and C. Guguen-Guillouzo. 2007. The human hepatoma
HepaRG cells: A highly differentiated model for studies of liver metabolism and toxicity of xenobiotics.
Chem. Biol. Interact. 168(1):66-73.
Halle, W. 2003. The Registry of Cytotoxicity: Toxicity testing in cell cultures to predict acute toxicity (LD50) and to
reduce testing in animals. ATLA 31(2):89-198.
Harris, T.W., N. Chen, F. Cunningham, M. Tello-Ruiz, I. Antoshechkin, C. Bastiani, T. Bieri, D. Blasiar, K.
Bradnam, J. Chan, C.K. Chan, W.J. Chen, P. Davis, E. Kenny, R. Kishore, D. Lawson, R. Lee, H.M. Muller,
C. Nakamura, P. Ozersky, A. Petcherski, A. Rogers, A. Sabo, E.M. Schwarz, K. Van Auken, Q. Wang, R.
Durbin, J. Spieth, P.W. Sternberg, and L.D. Stein. 2004. WormBase: A multi-species resource for nematode
biology and genomics. Nucleic Acids Res. 32(Database Issue):D411-D417.
Hentschel, D.M., K.M. Park, L. Cilenti, A.S. Zervos, I. Drummond, and J.V. Bonventre. 2005. Acute renal failure in
zebrafish: A novel system to study a complex disease. Am. J. Physiol. Renal. Physiol. 288(5):F923-F929.
Honegger, P., F. Monnet-Tschudi, and M.G. Zurich. 2009. Brain Aggregate Cell Cultures as an In Vitro Model for
Neurotoxicity Testing [online]. Available: http://alttox.org/brain-aggregate-cell-cultures-as-an-in-vitro-modelfor-neurotoxicity-testing/ [accessed March 19, 2015].
Houck, K.A., D.J. Dix, R.S. Judson, R.J. Kavlock, J. Yang, and E.L. Berg. 2009. Profiling bioactivity of the ToxCast chemical library using BioMAP primary human cell systems. J. Biomol. Screen. 14(9):1054-1066.
Howe, K, M.D. Clark, C.F. Torroja, J. Torrance, C. Berthelot, M. Muffato, J.E. Collins, S. Humphray, K. McLaren,
L. Matthews, S. McLaren, L. Sealy, M. Caccamo, C. Churcher, C. Scott, J.C. Barrett, R. Koch, G.J. Rauch, S.
White, W. Chow, B. Kilian, L.T. Quintais, J.A. Guerra-Assuno, Y. Zhou, Y. Gu, J. Yen, J.H. Vogel, T.
Eyre, S. Redmond, R. Banerjee, J. Chi, B. Fu, E. Langley, S.F. Maguire, G.K. Laird, D. Lloyd, E. Kenyon, S.
Donaldson, H. Sehra, J. Almeida-King, J. Loveland, S. Trevanion, M. Jones, M. Quail, D. Willey, A. Hunt, J.
Burton, S. Sims, K. McLay, B. Plumb, J. Davis, C. Clee, K. Oliver, R. Clark, C. Riddle, D. Elliot, G. Threadgold, G. Harden, D. Ware, S. Begum, B. Mortimore, G. Kerry, P. Heath, B. Phillimore, A. Tracey, N. Corby,
M. Dunn, C. Johnson, J. Wood, S. Clark, S. Pelan, G. Griffiths, M. Smith, R. Glithero, P. Howden, N. Barker,
C. Lloyd, C. Stevens, J. Harley, K. Holt, G. Panagiotidis, J. Lovell, H. Beasley, C. Henderson, D. Gordon, K.
Auger, D. Wrigh, J. Collins, C. Raisen, L. Dyer, K. Leung, I. Robertson, K. Ambridge, D. Leongamornlert, S.
McGuire, R. Gilderthorp, C. Griffiths, D. Manthravadi, S. Nichol, G. Barker, S. Whitehead, M. Kay, J.
Brown, C. Murnane, E. Gray, M. Humphries, N. Sycamore, D. Barker, D. Saunders, J. Wallis, A. Babbage, S.
Hammond, M. Mashreghi-Mohammadi, L. Barr, S. Martin, P. Wray, A. Ellington, N. Matthews, M. Ellwood,
R. Woodmansey, G. Clark, J. Cooper, A. Tromans, D. Grafham, C. Skuce, R. Pandian, R. Andrews, E. Harrison, A. Kimberley, J. Garnett, N. Fosker, R. Hall, P. Garner, D. Kelly, C. Bird, S. Palmer, I. Gehring, A. Berger, C.M. Dooley, Z. Ersan-rn, C. Eser, H. Geiger, M. Geisler, L. Karotki, A. Kirn, J. Konantz, M. Konantz, M. Oberlnder, S. Rudolph-Geiger, M. Teucke, C. Lanz, G. Raddatz, K. Osoegawa, B. Zhu, A. Rapp,
S. Widaa, C. Langford, F. Yang, S.C. Schuster, N.P. Carter, J. Harrow, Z. Ning, J. Herrero, S.M. Searle, A.
Enright, R. Geisler, R.H. Plasterk, C. Lee, M. Westerfield, P.J. de Jong, L.I. Zon, J.H. Postlethwait, C.
Nsslein-Volhard, T.J. Hubbard, H. Roest Crollius, J. Rogers, and D.L. Stemple. 2013. The zebrafish reference genome sequence and its relationship to the human genome. Nature 496(7446):498-503.
Huang, CJ, C.T. Tu, C.D. Hsiao, F.J. Hsieh, and H.J. Tsai. 2003. Germ-line transmission of a myocardium-specific
GFP transgene reveals critical regulatory elements in the cardiac myosin light chain 2 promoter of zebrafish.
Dev. Dyn. 228(1):30-40.
Huang, R., N. Southall, M.H. Cho, M. Xia, J. Inglese, and C.P. Austin. 2008. Characterization of diversity in toxicity mechanism using in vitro cytotoxicity assays in quantitative high throughput screening. Chem. Res. Toxicol. 21(3):659-667.
Prepublication Copy
71
Huang, R., M. Xia, M.H. Cho, S. Sakamuru, P. Shinn, K.A. Houck, D.J. Dix, R.S. Judson, K.l. Witt, R.J. Kavlock,
R.R. Tice, and C.P. Austin. 2011. Chemical genomics profiling of environmental chemical modulation of human nuclear receptors. Environ. Health Perspect. 119(8):1142-1148.
Huh, D., B.D. Matthews, A. Mammoto, M. Montoya-Zavala, H.Y. Hsin, and D.E. Ingber. 2010. Reconstituting organ-level lung functions on a chip. Science 328(5986):1662-1668.
Huh, D., D.C. Leslie, B.D. Matthews, J.P. Fraser, S. Jurek, G.A. Hamilton, K.S. Thorneloe, M.A. McAlexander, and
D.E. Ingber. 2012. A human disease model of drug toxicity-induced pulmonary edema in a lung-on-a-chip microdevice. Sci. Transl. Med. 4(159):159ra147.
Jennings, P. In press. The future of in vitro toxicology. Toxicology In Vitro.
Judson, R.S., K.A. Houck, R.J. Kavlock, T.B. Knudsen, M.T. Martin, H.M. Mortensen, D.M. Reif, D.M.Rotroff, I.
Shah, A.M. Richard, and D.J. Dix. 2010. In vitro screening of environmental chemicals for targeted testing
prioritization: The ToxCast project. Environ. Health Perspect. 118(4):485-492.
Judson, R., R. Kavlock, M. Martin, D. Reif, K. Houck, T. Knudsen, A. Richard, R.R. Tice, M. Whelan, M. Xia, R.
Huang, C. Austin, G. Daston, T. Hartung, J.R. Fowle III, W. Wooge, W. Tong, and D. Dix. 2013. Perspectives
on validation of high-throughput assays supporting 21st century toxicity testing. ALTEX 30(1):51-56.
Kalgutkar, A.S., and D. Dalvie. 2014. Predicting toxicities of reactive metabolite-positive drug candidates. Annu.
Rev. Pharmacol. Toxicol. 55:35-54.
Kanebratt, K.P., and T.B. Andersson. 2008. Evaluation of HepaRG cells as an in vitro model for human drug metabolism studies. Drug Metab. Dispos. 36(7):1444-1452.
Kavlock, R.J., C.P. Austin, and R.R. Tice. 2009. Toxicity testing in the 21st century: Implications for human health
risk assessment. Risk Anal. 29(4):485-487.
Kavlock, R.J., K. Chandler, K. Houck, S. Hunter, R. Judson, N. Kleinstreuer, T. Knudsen, M. Martin, S.Padilla, D.
Reif, A. Richard, D. Rotroff, N. Sipes, and D. Dix. 2012. Update on EPAs ToxCast program: Providing high
throughput decision support tools for chemical risk management. Chem. Res. Toxicol. 25(7):1287-1302.
Khetani, S.R., and S.N. Bhatia. 2008. Microscale culture of human liver cells for drug development. Nat. Biotechnol. 26(1):120-126.
Khetani, S.R., C. Kanchagar, O. Ukairo, S. Krzyzewski, A. Moore, J. Shi, S. Aoyama, M. Aleo, and Y. Will. 2013.
Use of micropatterned cocultures to detect compounds that cause drug-induced liver injury in humans. Toxicol. Sci. 132(1):107-117.
Kinsner-Ovaskainen, A., P. Prieto, S. Stanzel, and A. Kopp-Schneider. 2013. Selection of test methods to be included in a testing strategy to predict acute oral toxicity: An approach based on statistical analysis of data collected
in phase 1 of the ACuteTox project. Toxicol. In Vitro 27(4):1377-1394.
Klaper, R., D. Arndt, J. Bozich, and G. Dominguez. 2014. Molecular interactions of nanomaterials and organisms:
Defining biomarkers for toxicity and high-throughput screening using traditional and next-generation sequencing approaches. Analyst. 139(5):882-895.
Kleinstreuer, N.C., J. Yang, E.L. Berg, T.B. Knudsen, A.M. Richard, M.T. Martin, D.M. Reif, R.S. Judson, M. Polokoff, D.J. Dix, R.J. Kavlock, and K.A. Houck. 2014. Phenotypic screening of the ToxCast chemical library
to classify toxic and therapeutic mechanisms. Nat. Biotechnol. 32(6):583-591.
Knudsen, T.B., R.J. Kavlock, G.P. Daston, D. Stedman, M. Hixon, and J.H. Kim. 2011. Developmental toxicity testing for safety assessment: New approaches and technologies. Birth Defects Res. B Dev. Reprod. Toxicol.
92(5):413-420.
Kraushaar, U., T. Meyer, D. Hess, L. Gepstein, C.L. Mummery, S.R. Braam, and E. Guenther. 2012. Cardiac safety
pharmacology: From human ether-a-gogo related gene channel block towards induced pluripotent stem cell
based disease models. Expert Opin. Drug Saf. 11(2):285-298.
Krewski, D., M.E., Andersen, E. Mantus, and L. Zeise. 2009. Toxicity testing in the 21st century: Implications for
human health risk assessment. Risk Anal. 29(4):474-479.
Lamb, J., E.D. Crawford, D. Peck, J.W. Modell, I.C. Blat, M.J. Wrobel, J. Lerner, J.P. Brunet, A. Subramanian,
K.N. Ross, M. Reich, H. Hieronymus, G. Wei, S.A. Armstrong, S.J. Haggarty, P.A. Clemons, R. Wei, S.A.
Carr, E.S. Lander, and T.R. Golub. 2006. The Connectivity Map: Using gene-expression signatures to connect
small molecules, genes, and disease. Science 313(5795):1929-1935.
Langheinrich, U., G. Vacun, and T. Wagner. 2003. Zebrafish embryos express an orthologue of HERG and are sensitive toward a range of QT-prolonging drugs inducing severe arrhythmia. Toxicol. Appl. Pharmacol.
193(3):370-382.
Larson, B., P. Banks, S. Shultz, M. Sobol, and J.J. Cali. 2011. The utility of semi-automating multiplexed assays for
ADME/Tox applications. Comb. Chem. High Throughput Screen. 14(8):658-668.
72
Prepublication Copy

Leite, S.B., I. Wilk-Zasadna, J.M. Zaldivar, E. Airola, M.A. Reis-Fernandes, M. Mennecozzi, C. Guguen-Guillouzo,
C. Chesne, C. Guillou, P.A. Alves, and S. Coecke. 2012. Three-dimensional HepaRG model as an attractive
tool for toxicity testing. Toxicol. Sci. 130(1):106-116.
Legradi, J., N. El Abdellaoui, M. van Pomeren, and J. Legler. In press. Comparability of behavioural assays using
zebrafish larvae to assess neurotoxicity. Environ. Sci. Pollut. Res.
Lessman, C.A. 2011. The developing zebrafish (Danio rerio): A vertebrate model for high-throughput screening of
chemical libraries. Birth Defects Res. Part C Embryo Today 93(3):268-280.
Limbourg, B., and M. Zalokar. 1973. Permeabilization of Drosophila eggs. Dev. Biol. 35(2):382-387.
Lin, Z., and Y. Will. 2012. Evaluation of drugs with specific organ toxicities in organ-specific cell lines. Toxicol.
Sci. 126(1):114-127.
Love, D.R., F.B. Pichler, A. Dodd, B.R. Copp, and D.R. Greenwood. 2004. Technology for high-throughput
screens: The present and future using zebrafish. Curr. Opin. Biotechnol. 15(6):564-571.
Lubberstedt, M., U. Muller-Vieira, M. Mayer, K.M. Biemel, F. Knospel, D. Knobeloch, A.K. Nussler, J.C. Gerlach,
and K. Zeilinger. 2011. HepaRG human hepatic cell line utility as a surrogate for primary human hepatocytes
in drug metabolism assessment in vitro. J. Pharmacol. Toxicol. Methods 63(1):59-68.
MacArthur, R., W. Leister, H. Veith, P. Shinn, N. Southall, C.P. Austin, J. Inglese, and D.S. Auld. 2009. Monitoring
compound integrity with cytochrome P450 assays and qHTS. J. Biomol. Screen 14(5):538-546.
Marroquin, L.D., J. Hynes, J.A. Dykens, J.D. Jamieson, and Y. Will. 2007. Circumventing the Crabtree effect: Replacing media glucose with galactose increases susceptibility of HepG2 cells to mitochondrial toxicants. Toxicol. Sci. 97(2):539-547.
Maschmeyer, I., T. Hasenberg, A. Jaenicke, M. Lindner, A.K. Lorenz, J. Zech, L.A. Garbe, F. Sonntag, P. Hayden,
S. Ayehunie, R. Lauster, U. Marx, E.M. Materne. In press. Chip-based human liver-intestine and liver-skin cocultures - A first step toward systemic repeated dose substance testing in vitro. Eur. J. Pharm. Biopharm.
Mazzio, E.A., Y.I. Soliman, and K.F. Soliman. 2010. Variable toxicological response to the loss of OXPHOS
through 1-methyl-4-phenylpyridinium-induced mitochondrial damage and anoxia in diverse neural immortal
cell lines. Cell Biol. Toxicol. 26(6):527-539.
Mathur, A., P. Loskill, S. Hong, J. Lee, S.G. Marcus, L. Dumont, B.R. Conklin, H. Willenbring, L.P. Lee, and K.E.
Healy. 2013. Human induced pluripotent stem cell-based microphysiological tissue models of myocardium
and liver for drug development. Stem Cell Res. Ther. 4(Suppl. 1):S14.
McAlister, G.C., D.P. Nusinow, M.P. Jedrychowski, M. Whr, E.L. Huttlin, B.K. Erickson, R. Rad, W. Haas, and
S.P. Gygi. 2014. MultiNotch MS3 enables accurate, sensitive, and multiplexed detection of differential expression across cancer cell line proteomes. Anal. Chem. 86(14):7150-7158.
McVey, K.A., J.A. Mink, I.B. Snapp, W.S. Timberlake, C.E. Todt, R. Negga, and V.A. Fitsanakis. 2012. Caenorhabditis elegans: An emerging model system for pesticide neurotoxicity. J. Environ. Anal. Toxicol. 2:125,
doi: 10.4172/2161-0525.S4-003.
Melstrom, P.C., and P.L. Williams. 2007. Reversible AChE inhibitors in C. elegans vs. rats, mice. Biochem. Biophys. Res. Commun. 357(1):200-205.
Messner, S., I. Agarkova, W. Moritz, and J.M. Kelm. 2013. Multi-cell type human liver in microtissues for hepatotoxicity testing. Arch. Toxicol. 87(1):209-213.
Meyer, D., and P.L. Williams. 2014. Toxicity testing of neurotoxic pesticides in Caenorhabditis elegans. J. Toxicol.
Environ. Health B Crit. Rev. 17(5):284-306.
Milan, D.J., T.A. Peterson, J.N. Ruskin, R.T. Peterson, and C.A. MacRae. 2003. Drugs that induce repolarization
abnormalities cause bradycardia in zebrafish. Circulation.107(10):1355-1358.
NAS (National Academy of Sciences). 2014. The Potential of the Tissue Chip for Environmental Health Studies.
Emerging Science For Environmental Health Decisions Newsletter Issue II [online]. Available: http://nassites.org/emergingscience/files/2013/05/tissue-chips-final.pdf [accessed March 20, 2015].
Nichols, C.D. 2006. Drosophila melanogaster neurobiology, neuropharmacology, and how the fly can inform central
nervous system drug discovery. Pharmacol. Ther. 112(3):677-700.
NIEHS (National Institute of Environmental Health Sciences). 2001a. Report of the International Workshop on In
Vitro Methods for Assessing Acute Systemic Toxicity. NIH Publication 01-4499. National Toxicology Program [online]. Available: http://ntp.niehs.nih.gov/pubhealth/evalatm/test-method-evaluations/acute-systemictox/in-vitro/wksp-rpt-recs/index.html#Complete-Report [accessed March 20, 2015].
NIEHS (National Institute of Environmental Health Sciences). 2001b. Guidance Document on Using In Vitro Data
to Estimate In Vivo Starting Doses for Acute Toxicity Based on Recommendations from an International
Workshop Organized by ICCVAM and NICEA TM. NIH Publication 01-4500. National Toxicology Program
Prepublication Copy
73
[online]. Available: http://ntp.niehs.nih.gov/pubhealth/evalatm/test-method-evaluations/acute-systemic-tox/invitro/guidance-document/index.html [accessed March 20, 2015].
NRC (National Research Council). 2007. Toxicity Testing in the 21st Century. Washington, DC: The National
Academies Press.
NTP (National Toxicology Program). 2004. A National Toxicology Program for the 21st Century: A Roadmap for
the Future [online]. Available: https://ntp.niehs.nih.gov/ntp/about_ntp/ntpvision/ntproadmap_508.pdf [accessed March 23, 2015].
OBrien, P. 2014. High-content analysis in toxicology: Screening substances for human toxicity potential, elucidating subcellular mechanisms and in vivo use as translational safety biomarkers. Basic Clin. Pharmacol. Toxicol. 115(1):4-17.
O'Brien, P.J., W. Irwin, D. Diaz, E. Howard-Cofield, C.M. Krejsa, M.R. Slaughter, B. Gao, N. Kaludercic, A. Angeline, P. Bernardi, P. Brain, and C. Hougham. 2006. High concordance of drug-induced human hepatotoxicity
with in vitro cytotoxicity measured in a novel cell-based model using high content screening. Arch. Toxicol.
80(9):580-604.
OECD (Organization for Economic Co-operation and Development). 2013a. Test No. 210: Fish, Early-life Stage
Toxicity Test. OECD Guidelines for the Testing of Chemicals, Section 2. Paris, France: OECD Publishing.
OECD (Organization for Economic Co-operation and Development). 2013b. Test No. 236: Fish Embryo Acute Toxicity (FET) Test. OECD Guidelines for the Testing of Chemicals, Section 2. Paris, France: OECD Publishing.
OECD (Organization for Economic Co-operation and Development). 2014. Guidance Document for Describing
Non-guideline In Vitro Test Methods. Series on Testing and Assessment No. 211 [online]. Available:
http://www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=ENV/JM/MONO(2014)35&doclang
uage=en [accessed March 24, 2015].
Pamies, D., T. Hartung, and H.T. Hogberg. 2014. Biological and medical applications of a brain-on-a-chip. Exp.
Biol. Med. 239(9):1096-1107.
Patlewicz, G., T. Simon, K. Goyak, R.D. Phillips, J.C. Rowlands, S.D. Seidel, and R.A. Becker. 2013. Use and validation of HT/HC assays to support 21st century toxicity evaluations. Regul. Toxicol. Pharmacol. 65(2):259268.
Perkins, E.J., G.T. Ankley, K.M. Crofton, N. Garcia-Reyero, C.A. LaLone, M.S. Johnson, J.E. Tietge and D.L.
Villeneuve. 2013. Current perspectives on the use of alternative species in human health and ecological hazard
assessments. Environ. Health Perspect. 121(9):1002-1010.
Persson, M., A. Lye, M. Jacquet, N. Mow, A. Thougaard, T. Mow, and J. Hornberg. 2014. High-content analysis/screening for predictive toxicology: Application to hepatotoxicity and genotoxicity. Basic Clin. Pharmacol.
Toxicol. 115(1):18-23.
Peterson, R.T., R. Nass, W.A. Boyd, J.H. Freedman, K. Dong, and T. Narahashi. 2008. Use of non-mammalian alternative models for neurotoxicological study. Neurotoxicology 29(3):546-555.
Pfannenbecker, U., S. Bessou-Touya, C. Faller, J. Harbell, T. Jacob, H. Raabe, M. Tailhardat, N. Alepee, A. De
Smedt, B. De Wever, P. Jones, Y. Kaluzhny, B. Le Varlet, P. McNamee, M. Marrec-Fairley, and F. Van Goethem. 2013. Cosmetics Europe multi-laboratory pre-validation of the epiocular reconstituted human tissue test
method for the prediction of eye irritation. Toxicol. In Vitro 27(2):619-626.
Podratz, J.L., N.P. Staff, D. Froemel, A. Wallner, F. Wabnig, A.J. Bieber, A. Tang, and A.J. Windebank. 2011. Drosophila melanogaster: A new model to study cisplatin-induced neurotoxicity. Neurobiol. Dis. 43(2):330-337.
Podratz, J.L., N.P. Staff, J.B. Boesche, N.J. Giorno, M.E. Hainy, S.A. Herring, M.T. Klennert, C. Milaster, S.E.
Nowakowski, R.G. Krug II, Y. Peng, and A.J. Windebank. 2013. An automated climbing apparatus to measure chemotherapy-induced neurotoxicity in Drosophila melanogaster. Fly (Austin) 7(3):187-192.
Porceddu, M., N. Buron, C. Roussel, G. Labbe, B. Fromenty, and A. Borgne-Sanchez. 2012. Prediction of liver injury induced by chemicals in human with a multiparametric assay on isolated mouse liver mitochondria. Toxicol. Sci. 129(2):332-345.
Potts, S.J., T.D. Johnson, F.A. Voelker, H. Lange, and G.D. Young. 2011. Multiplexed measurement of proteins in
tissue in a clinical environment. Appl. Immunohistochem. Mol. Morphol. 19(6):494-498.
Ralda, D., and B. Pia. 2014. In vivo zebrafish assays for analyzing drug toxicity. Expert Opin. Drug Metab. Toxicol. 10(5):685-697.
Rand, M.D. 2010. Drosophotoxicology: The growing potential for Drosophila in neurotoxicology. Neurotoxicol.
Teratol. 32(1):74-83.
Ravindranath, A.C., N. Perualila-Tan, A. Kasim, G. Drakakis, S. Liggi, S.C. Brewerton, D. Mason, M.J. Bodkin,
D.A. Evans, A. Bhagwat, W. Talloen, H.W. Ghlmann, Z. Shkedy, and A. Bender. 2015. Connecting gene
74
Prepublication Copy

expression data from connectivity map and in silico target predictions for small molecule mechanism-ofaction analysis. Mol. BioSyst. 11(1):86-96.
Ruan, Q.L., J.J. Ju, Y.H. Li, R. Liu, Y.P. Pu, L.H. Yin, and D.Y. Wang. 2009. Evaluation of pesticide toxicities with
differing mechanisms using Caenorhabditis elegans. J. Toxicol. Environ. Health A. 72(11-12):746-751.
Rubin, G.M., and E.B. Lewis. 2000. A brief history of Drosophila's contributions to genome research. Science
24(5461):2216-2218.
Sauer. U.G., S. Vogel, A. Hess, S.N. Kolle, L. Ma-Hock, B. van Ravenzwaay, and R. Landsiedel. 2013. In vivo-in
vitro comparison of acute respiratory tract toxicity using human 3D airway epithelial models and human A549
and murine 3T3 monolayer cell systems. Toxicol. In Vitro 27(1):174-190.
Sauer, U.G., A. Aumann, L. Ma-Hock, R. Landsiedel, and W. Wohlleben. 2015. Influence of dispersive agent on
nanomaterial agglomeration and implications for biological effects in vivo or in vitro. Toxicol. In Vitro
29(1):182-186.
Sawin-McCormack, E.P., M.B. Sokolowski, and A.R. Campos. 1995. Characterization and genetic analysis of Drosophila melanogaster photobehavior during larval development. J. Neurogenet. 10(2):119-135.
Segalat, L. 2007. Invertebrate animal models of diseases as screening tools in drug discovery. ACS Chem. Biol.
2(4):231-236.
Shukla, S.J., R. Huang, C.P. Austin, and M. Xia. 2010. The future of toxicity testing: A focus on in vitro methods
using a quantitative high throughput screening platform. Drug Discov. Today 15(23-24):997-1007.
Simoneschi, D., F. Simoneschi, and N.E. Todd. 2014. Assessment of cardiotoxicity and effects of malathion on the
early development of zebrafish (Danio rerio) using computer vision for heart rate quantification. Zebrafish
11(3):275-280.
Sipes, N.S., M.T. Martin, P. Kothiya, D.M. Reif, R.S. Judson, A.M. Richard, K.A. Houck, D.J. Dix, R.J. Kavlock,
and T.B. Knudsen. 2013. Profiling 976 ToxCast chemicals across 331 enzymatic and receptor signaling assays. Chem. Res. Toxicol. 26(6):878-895.
Sokolowski, M.B. 2001. Drosophila: Genetics meets behaviour. Nat. Rev. Genet. 2(11):879-890.
Spielmann, H., E. Genschow, M. Leibsch, and W. Halle. 1999. Determination of the starting dose for acute oral toxicity (LD50) testing in the up and down procedure (UDP) from cytotoxicity data. ATLA 27(6):957-966.
Spielmann, H., S. Hoffmann, M. Liebsch, P. Botham, J.H. Fentem, C. Eskes, R. Roguet, J. Cotovio, T. Cole, A.
Worth, J. Heylings, P. Jones, C. Robles, H. Kandarova, A. Gamer, M. Remmele, R. Curren, H. Raabe, A.
Cockshott, I. Gerner, and V. Zuang. 2007. The ECVAM international validation study on in vitro tests for
acute skin irritation: Report on the validity of the EPISKIN and EpiDerm assays and on the Skin Integrity
Function Test. ATLA 35(6):559-601.
Steinritz, D., N. Mhle, C. Pohl, M. Papritz, B. Stenger, A. Schmidt, C.J. Kirkpatrick, H. Thiermann, R. Vogel, S.
Hoffmann, and M. Aufderheide. 2013. Use of the Cultex Radial Flow System as an in vitro exposure method
to assess acute pulmonary toxicity of fine dusts and nanoparticles with special focus on the intra- and interlaboratory reproducibility. Chem. Biol. Interact. 206(3):479-490.
Sung, J.H., B. Srinivasan, M.B. Esch, W.T. McLamb, C. Bernabini, M.L. Shuler, and J.J. Hickman. 2014. Using
physiologically-based pharmacokinetic-guided body-on-a-chip systems to predict mammalian response to
drug and chemical exposure. Exp. Biol. Med. (Maywood) 239(9):1225-1239.
Takahashi, K., and S. Yamanaka. 2006. Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell 126 (4):663-676.
Taylor, K.L., N.J. Grant, N.D. Temperley, and E.E. Patton. 2010. Small molecule screening in zebrafish: An in vivo
approach to identifying new chemical tools and drug leads. Cell Commun. Signal 8:11.
Thomas, R.S., M.B. Black, L. Li, E. Healy, T.M. Chu, W. Bao, M.E. Andersen, and R.D. Wolfinger. 2012. A comprehensive statistical analysis of predicting in vivo hazard using high-throughput in vitro screening. Toxicol.
Sci. 128(2):398-417.
Tsang, M. 2010. Zebrafish: A tool for chemical screens. Birth Defects Res. Part C Embryo Today 90(3):185-192.
Villatte, F., V. Marcel, S. Estrada-Mondaca, and D. Fournier. 1998. Engineering sensitive acetylcholinesterase for
detection of organophosphate and carbamate insecticides. Biosens. Bioelectron. 13(2):157-164.
Vliegenthart, A.D., C.S. Tucker, J. Del Pozo, and J.W. Dear. 2014. Zebrafish as model organisms for studying drug
induced liver injury. Br. J. Clin. Pharmacol. 78(6):1217-1227.
Whitworth, A.J., P.D. Wes, and L.J. Pallanck. 2006. Drosophila models pioneer a new approach to drug discovery
for Parkinson's disease. Drug Discov. Today 11(3-4):119-126.
Will, Y., and J. Dykens. 2014. Mitochondrial toxicity assessment in industrya decade of technology development
and insight. Expert Opin. Drug Metab. Toxicol. 10(8):1061-1067.
Prepublication Copy
75
Williams, P.L., and D.B. Dusenbery. 1990. A promising indicator of neurobehavioral toxicity using the nematode
Caenorhabditis elegans and computer tracking. Toxicol. Ind. Health 6(3-4):425-440.
Wilmes, A., A. Limonciel, L. Aschauer, K. Moenks, C. Bielow, M.O. Leonard, J. Hamon, D. Carpi, S. Ruzek, A.
Handler, O. Schmal, K. Herrgen, P. Bellwon, C. Burek, G.L. Truisi, P. Hewitt, E. Di Consiglio, E. Testai, B.J.
Blaauboer, C. Guillou, C.G. Huber, A. Lukas, W. Pfaller, S.O. Mueller, F.Y. Bois, W. Dekant, and P. Jennings. 2013. Application of integrated transcriptomic, proteomic and metabolomic profiling for the delineation
of mechanisms of drug induced cell stress. J. Proteomics. 79:180-194.
Wilmes, A., C. Bielow, C. Ranninger, P. Bellwon, L. Aschauer, A. Limonciel, H. Chassaigne, T. Kristl, S. Aiche,
C.G. Huber, C. Guillou, P. Hewitt, M.O. Leonard, W. Dekant, F. Bois, and P. Jennings. In press. Mechanism
of cisplatin proximal tubule toxicity revealed by integrating transcriptomics, proteomics, metabolomics and
biokinetics. Toxicology In Vitro.
Wink, S., S. Hiemstra, S. Huppelschoten, E. Danen, M. Niemeijer, G. Hendriks, H. Vrieling, B. Herpers, and B. van
de Water. 2014. Quantitative high content imaging of cellular adaptive stress response pathways in toxicity for
chemical safety assessment. Chem. Res. Toxicol. 27(3):338-355.
Wu, T.S., J.J. Yang, F.Y. Yu, and B.H. Liu. 2012. Evaluation of nephrotoxic effects of mycotoxins, citrinin and patulin, on zebrafish (Danio rerio) embryos. Food Chem. Toxicol. 50(12):4398-4404.
Xia, M., R. Huang, K.L. Witt, N. Southall, J. Fostel, M.H. Cho, A. Jadhav, C.S. Smith, J. Inglese, C.J. Portier, R.R.
Tice, and C.P. Austin. 2008. Compound cytotoxicity profiling using quantitative high-throughput screening.
Environ. Health Perspect. 116(3):284-291.
Xu, J.J., P.V. Henstock, M.C. Dunn, A.R. Smith, J.R. Chabot, and D. de Graaf. 2008. Cellular imaging predictions
of clinical drug-induced liver injury. Toxicol. Sci. 105(1):97-105.
Yamamoto, N., K. Komori, K. Montagne, H. Matsui, H. Nakayama, S. Takeuchi, and Y. Sakai. 2011. Cytotoxicity
evaluation of reactive metabolites using rat liver homogenate microsome-encapsulated alginate gel microbeads. J. Biosci. Bioenerg. 111(4):454-458.
Yen, J., S. Donerly, E.D. Levin, and E.A. Linney. 2011. Differential acetylcholinesterase inhibition of chlorpyrifos,
diazinon and parathion in larval zebrafish. Neurotoxicol. Teratol. 33(6):735-741.
76
Prepublication Copy
5
Integration and Decision-Making for Predictive Toxicology
As described in Chapter 2, a robust integration and decision-making strategy is needed as part of the
overall approach developed by the committee to predict acute, debilitating toxicity. Specifically, this
chapter will describe general approaches and considerations for integrating data to make the categorization decisions outlined in the tiered prioritization strategy described in Chapter 2. Discussions of these
topics are related to the task of evaluating chemicals for their potential to elicit acute toxicity. The committee has also noted a number of topics beyond its charge on which the Department of Defense (DOD)
will need to make policy decisions in light of specific needs.
GENERAL APPROACH TO INTEGRATION AND DECISION-MAKING
Integration and decision-making to support prediction of the potential of chemicals to cause acute toxicity are needed at many levels. As described in Chapter 2, the goal at each tier of the prioritization strategy
is to place chemicals in three categories: high confidence of high toxicity, high confidence of low toxicity, and inadequate data. Box 5-1 presents a simplified illustration of the process to base decisions on the
results of a single model for a single end point. As illustrated in this simple case, categorization depends on
defining clear benchmarks that set the boundaries for high and low toxicity and on taking uncertainty or
confidence in each individual prediction into account. Key tasks for DOD will be determining the appropriate benchmarks for each end point that is relevant to the evaluation of acute toxicity and specifying the level
of confidence appropriate to its needs.
In the more general case in which there are several predictions for multiple end points, the committee divided the integration and decision-making process into two parts, as illustrated in Figure 5-1:
Withinend point integration and decision-making, which is based on integrating various data
streams and predictions that inform a single acute-toxicity end point.
Crossend point integration and decision-making, which is based on integrating predictions
from several acute-toxicity end points.
WithinEnd Point Integration and Decision-making
As discussed in Chapter 2, concern about potential acute toxicity spans a wide range of chemicals in
terms of structures and physicochemical properties. Moreover, as shown in Chapters 3 and 4 and by others (Bauer-Mehren et al. 2012), no single prediction approach, whether a nontesting approach or an assaybased approach, is sufficient to capture the entire chemical domain. For some end points, such as a rat
LD50, multiple models and tools that have various degrees of accuracy are available (see, for example,
review by Diaza et al. 2015), and it might be of interest to integrate their predictions into a single summary prediction. Furthermore, a chemicals pharmacokineticsabsorption, distribution, metabolism, and
excretion (ADME)might contribute to its potency and toxicity. Therefore, even within an end point
domain, there might be a need to integrate multiple databases, assays, models, and tools to develop an
Prepublication Copy
77
integrated prediction for that end point. The committee notes that using various integrative approaches
might also help to identify biological responses that can explain chemical-induced adverse reactions
(Bauer-Mehren et al. 2012).
A key task for DOD will be defining the most informative end point domains for its application,
for example, whether to define an end point at a more general level, such as neurotoxicity, or at a more
specific level, such as seizure or cholinesterase inhibition.
BOX 5-1 Simplified Illustration of Integration and Decision-making

Models and End Points: At Tier 1, a single quantitative structureactivity relationship (QSAR) model
is being used to predict rat LD50 values from chemical structure and physicochemical properties. No
other models or end points are being considered. To illustrate a simplified approach, chemicals outside a specified applicability domain are placed in the inadequate data category for this example.
For chemicals inside the applicability domain, the output of the model is an LD50 estimate with a confidence interval that reflects uncertainty.
Chemicalstructure,
LD50prediction
QSAR
physicochemical
andconfidence
Integration: Because only a single model
model
properties
interval
and a single end point are being considered,
there is no integration of different predicOutsideapplicability
tions.
domain
Inadequatedata
Category Benchmarks: A set of reference
chemicals based on DOD interests that have known (possibly more than one) LD50 values are selected to represent the high toxicity and low toxicity categories from which category benchmarks are
derived. For example, the high toxicity benchmark could be defined as the highest LD50 of the least
toxic high toxicity reference chemical. For some end points, generic toxicity benchmarks are available, such as the European Union and Global Harmonized System categories for acute toxicity.
Quantile
LD50Prediction
Highconfidenceof
lowtoxicity
Predicted
LD50s
High()and
low()toxicity
benchmarksbased
onreference
compoundswith
knownLD50s
mg/kg
Inadequatedata
Highconfidenceof
hightoxicity
Highconfidenceof
Highconfidence
Inadequatedata
ofhightoxicity
lowtoxicity
Decision-making: Decisions as to how to place chemicals into categories are based on the confidence bounds for each prediction:
A chemical is categorized as high confidence of high toxicity if the upper confidence bound on
the predicted LD50 is less than or equal to the high toxicity benchmark.
A chemical is categorized as high confidence of low toxicity if the lower confidence bound on
the predicted LD50 is greater than or equal to the low toxicity benchmark.
The remaining chemicals are categorized as having inadequate data.
78
Prepublication Copy
WithinEndPointIntegration
andDecisionMaking
ChemicalStructure,
Physicochemical
Properties
BiologicalActivityin
CellsandLower
Organisms
MultipleDatabases,Assays,
Models,andTools
CrossEndPointIntegration
andDecisionMaking
AcuteToxicityScorecard
Endpoint
domain
High
confidence
ofhigh
toxicity
Toxicity
Benchmarks
Inadequate
data
Death
(LD50)
Neuro
toxicity
IntegratedPrediction
High
confidence
oflow
toxicity
Pulmonary
toxicity
...
Quantile
PulmonaryToxicity
Neurotoxicity
Death(LD50)
Highconfidenceof
hightoxicity
Highconfidenceof
lowtoxicity
Predicted
LD50
Reference
compounds
Inadequatedata
mg/kg
Highconfidence
Highconfidenceof
Inadequatedata
lowtoxicity
ofhightoxicity
ToxPi visualization
withcategorybenchmarks
FIGURE 5-1 Illustration of a general approach to integration and decision-making for applying predictive approaches to acute, debilitating toxicity. Withinend point integration and decision-making has three basic steps:
(1) developing an integrated prediction from potentially multiple databases, assays, models, and tools; (2) specifying
toxicity benchmarks; and (3) placing chemicals into the appropriate category for the end point under consideration.
Crossend point integration and decision-making could consist of (1) a simple scorecard for a chemical in
which each individual end pointspecific decision is recorded or (2) more integrative approaches, such as ToxPi,
that include the underlying toxicity end point predictions from which categories were assigned. The example presented here illustrates how information must be able to translate between withinend point and crossend point integration.
There are three steps in reaching a decision about a particular toxicity end point (the last two are the
same as in the simple case described previously):
(1) Integrating potentially multiple databases, assays, models, and tools into an integrated prediction, with its confidence interval, as to a chemicals toxicity potential for that end point.
(2) Specifying toxicity benchmarks that define the thresholds for what is considered high or low
toxicity for that end point.
(3) Placing chemicals into categories on the basis of the specified toxicity benchmarks, taking into
account the confidence interval of the integrated prediction.
Some of the available methods for each step are described in greater detail below.
Prepublication Copy
79
CrossEnd Point Integration and Decision-Making
As described in Chapter 2, the concern about potential acute toxicity spans a wide array of toxic end
points and biological mechanisms. From a military perspective, any acute toxicity that is severe enough to
cause debilitation or death is enough to warrant concern. Thus, a simple approach to crossend point integration would be simply to summarize the categorization results for each end point in a single scorecard,
as shown in the upper right of Figure 5-1. Each end point would be evaluated as described above as to
whether for a given type of acute toxicity (such as neurotoxicity) the chemical exhibited high toxicity,
low toxicity, or inadequate data. Then, in integrating into an overall evaluation, a chemical will be
sorted into the high toxicity bin if at least one of the end points is high toxicity, the low toxicity bin
if all the end points are low toxicity, and the inadequate data bin if neither of the first two conditions
applies. That approach also has the advantage of retaining the end pointspecific information, so that future data generation can be targeted better. The approach is also consistent with a low tolerance for false
negatives in that each end point identified as predictive of an acute toxic (debilitating or lethal) response
serves as sufficient evidence to place a chemical into a high toxicity bin.
Crossend point integration can also be visualized in a recombination approach (such as ToxPi, discussed in more detail below) and perhaps even augmented with information on toxicity benchmarks as
illustrated in Figure 5-1. The recombination approach, however, suggests an alternative integration that
would not depend strictly on a simple decision rule related to the categories for each end point. For example, the ToxPi approach (see lower right of Figure 5-1) could also provide a summary measure that consists of a weighted sum of individual toxicity end points. Thus, even if each individual end point is rated
as inadequate, it is conceivable that the presence of multiple end points close to their benchmark
thresholds would allow the chemical to be categorized as high or low on the basis of the summary
measure even if no individual end point were so rated. Setting up appropriate decision rules would be a
key policy question for DOD if it chose to implement the committees suggested approach for predicting
acute, debilitating toxicity.
APPROACHES TO INTEGRATION
The general approaches for integrating databases, assays, models, and tools are illustrated in Figure
5-2 with LD50 as an example end point. Broadly, they can be divided into approaches that combine individual predictions into an integrated prediction (upper panel, A) and approaches that first combine the
underlying data from databases and assays before building an integrated model or tool (lower panel, B).
Specific approaches are described in more detail below, particularly in relation to predicting acute toxicity; examples of applications to nontesting approaches (Chapter 3), biological assay-based approaches
(Chapter 4), and combinations of the two are provided if possible.
Meta-Analytic Approaches
From a formal statistical perspective, the rich literature on meta-analysis offers guidance on how to
aggregate information from multiple independent sources (Borenstein et al. 2009). It is expected that appropriate meta-analysis will help to improve the results from individual studies that might have been underpowered or have suffered from noise, bias, and absence of data. Meta-analysis can also help to reveal
interesting patterns or relationships among studies and generate results that are statistically robust. The
committee did not locate any examples of published meta-analyses of acute-toxicity predictions built
from the types of data described in Chapters 3 and 4. However, for such an end point as LD50, for which
multiple tools and models are available in the same chemical domain (for example, the five models reviewed by Diaza et al. 2015), a meta-analytic approach might be considered to combine results. In addition to combining individual predictions, meta-analytic approaches could be applied to individual categorization decisions (for example, see Box 5-2).
80
Prepublication Copy
ChemicalStructure,
Physicochemical
Properties
QSAR
model
LD50prediction
andconfidence
interval
Biochemical
Properties,Biological
ActivityinCellsand
LowerOrganisms
Biological
model
LD50prediction
andconfidence
interval
IntegratedLD50
predictionand
confidence
interval
Metaanalysis(e.g.,randomeffects)
Recombination(e.g.,ToxPi)
B
ChemicalStructure,
Physicochemical
Properties
Biochemical
Properties,Biological
ActivityinCellsand
LowerOrganisms
Combined
physicochemical
andbiologicaldata
Integrated
model
IntegratedLD50
predictionand
confidence
interval
Pooleddatamodels(e.g.,machinelearning)
Sequentialmodels(e.g.,reversetoxicokinetics)
Hierarchicalmodels(e.g.,chemicalbiologicalreadacross)
FIGURE 5-2 Approaches for integrating disparate datasets with LD50 as an example. A: approaches that keep datasets separate (such as physicochemical data and assay-based biological-activity data) and integrate predictions
from models developed for each dataset independently. B: approaches that combine datasets before modeling and
develop a new integrated model that makes a prediction from the combined dataset.
Three key issues should be considered in conducting any meta-analyses, including those applied to
acute-toxicity end points. First, individual results should be investigated to determine whether the data
can be combined reliably (Crowther et al. 2010). For example, results that are to be combined should be
associated with a common predicted acute-toxicity end point. Often, decisions on which statistical techniques to use for meta-analysis are not as important as decisions on which studies are to be combined because later analysis will not be able to correct for an inappropriate combination of studies. Second, an effective meta-analysis should ensure that better results receive more weight during information
aggregation (Crowther et al. 2010). For example, it is reasonable for predictions that have less variance or
that are based on a larger sample size to contribute more heavily to the overall summary statistic in a meta-analysis. Other factors, such as biases and methodological strengths and weaknesses of individual approaches, can also be included, either qualitatively or quantitatively. Third, several statistical methods can
be used to combine results. They include methods that are based on results of significance testing, such as
p values or z scores, and fixed and random effect models that use summary statistics, such as the mean
and standard error derived from individual results (Hedges and Olkin 1985; Borenstein et al. 2009). For
all three issues, sensitivity analysis can be performed to assess the effects of various choices on the results
of a meta-analysis (Higgins and Green 2011).
Prepublication Copy
81
BOX 5-2 Example of Meta-Analytic Approach That Uses

Irreproducible Discovery Rates to Integrate Categorization Decisions
Meta-analytic approaches, such as the irreproducible discovery rate (IDR) (Li et al. 2011), can be
used to measure the reproducibility of results from two independent studies and to filter noise in the
results. In the example below, two studies (Study I and Study II) have made toxicity predictions, and
the results are categorized into high toxicity, inadequate data, or low toxicity, on the basis of simulated category benchmarks with thresholds at values of [010], [1060], and [60100], respectively.
However, many chemicals show inconsistent categorizations between the two studies. To integrate
the two results, IDR considers the reproducibility between studies to assign a chemical-specific reproducibility measure. The far-right column presents the reproducibility measure as 1 local IDR, where
the chemicals that have the highest values are highlighted in red. Note that taking reproducibility
among studies into account can change the categorization of high toxicity.
Chemicals
CategoryBenchmarksbyinitial
toxicitypredictionscores:
Hightoxicity:
[60100]
Inadequatedata:
(1060)
Lowtoxicity:
[010]
Theonescategorizedashigh
toxicitybyinitialpredictionscores
arehighlightedinBlue
Theoneswithtop6(1localIDR)
scores(orthe6mostreproducible
results)arehighlightedinRed
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
InitialToxicityPredictionScores
StudyI
StudyII
99
100
72
45
70
58
66
0
65
38
62
49
56
15
51
60
51
72
47
59
40
53
37
42
36
31
30
20
30
20
29
35
24
28
15
28
14
43
8
41
1 localIDR
0.99
0.53
0.91
0.00
0.19
0.59
0.00
0.62
0.79
0.45
0.17
0.06
0.03
0.01
0.01
0.02
0.01
0.00
0.00
0.00
Recombination-Based Approaches
Several approaches for data integration have recently been developed to handle new problems presented by high-dimensional toxicity data, such as information from different biological assays. Because
the results from different data streams are not strictly comparable, formal meta-analytic approaches are
not immediately applicable. There are a variety of methods for weighting multiple streams of evidence
differently, from largely qualitative, expert-judgment approaches (for example, Hill criteria) to quantitative statistical frameworks that formalize weighting schemes (Linkov et al. 2015). Given the need to categorize chemicals, intermediate approaches that are quantitative and incorporate expert judgment are likely
to be most useful in predicting acute toxicity. The Toxicological Prioritization Index (ToxPi) developed
by Reif et al. (2010) provides a useful illustration of how to combine multiple data streams (discussed at
length in NRC 2014).
82
Prepublication Copy

The ToxPi combines data streams (physicochemical data, biological assays, or both) into a relative
index to facilitate prioritization or categorization. At its most basic, a summary ranking is derived for each
chemical on the basis of a weighted sum of rankings for different data sources. Although advanced statistical approaches could be used to group and weight individual pieces of evidence empirically, published
applications have used substantial expert judgment and taken into account the specific sources of data
being integrated and the context of integration. For example, binding assays for several cytochrome P450
(CYP) isoforms might be run to assess xenobiotic metabolism, but isoforms expressed in a target tissue
of interest might warrant more weight. Thus, using the ToxPi approach and reference chemicals can provide an overall, integrated ranking that can be used for categorization decisions (see Figure 5-3). In addition, the ToxPi provides a visualization of the individual component ranks and so can be used to support a
multicriteria decision-making scenario (Pavan and Worth 2008) in which categorization decisions involve integration of multiple, possibly conflicting criteria. For example, placing chemicals in the high
confidence of high toxicity bin could necessitate synthesizing results when individual pieces of evidence
serve as flags of high alert (such as solid evidence from a single assay that is deemed highly predictive of
acute toxicity). The use of ToxPi for crossend point integration is illustrated in the lower right of Figure
5-1, where axes are flagged if activity surpasses a benchmark threshold. ToxPi can also be applied to
withinend point integration.
Pooled Data-Based Approaches
The basic idea behind pooled data-based approaches is the use of the same types of nontesting approaches (such as read-across and QSAR) to make chemical-based or biologically based predictions.
Thus, the biological assay results are simply treated as additional biological descriptors that can be used
with physicochemical descriptors in building quantitative models. Several examples of this approach applied to acute toxicity are described in Chapter 3 (Lee et al. 2010; Sedykh et al. 2011); they retrospectively apply statistical methods, such as k-nearest neighbor, random forest, or multiple linear regression to a
combined dataset of chemical-structure data and literature-derived cytotoxicity data. The same approach
could be used more prospectively, in which new biological activity data generated in Tier 2 are combined
with chemical-structure information used previously in Tier 1 to develop an integrated prediction based
on pooling of both datasets.
Hierarchical Modeling Approaches
Hierarchical approaches can build-in information on chemical structure or global performance to inform modeling decisions (Wilson et al. 2014). For example, as noted in Chapter 3, several groups have
used biological data to stratify chemicals into clusters; more localized modeling, such as the use of
QSARs, was then applied to each (Zhu et al. 2009; Zhang 2011; Lounkine et al. 2012). Zhu et al. (2009)
applied such an approach to acute toxicity. Specifically, they grouped chemicals into those with and without good correlation between in vitro cytotoxicity IC50s and in vivo rodent LD50s. They then built one
QSAR model to assign chemicals to each group and two additional QSAR models to predict LD50s for
each group. They also compared their approach with the commercial TOPKAT software, using a set of
chemicals that was outside the training set of both approaches. They found that their two-step hierarchy of
QSAR models had a greater correlation coefficient and smaller mean absolute error. That type of stratification of QSAR models with biological information might be a fruitful integration approach that can be
tried with other acute-toxicity end points.
Sequential Modeling Approaches
Another approach to model-based integration is to connect models sequentially, that is, by using
outputs from one model as part of the inputs into another model. Such an approach can be developed
Prepublication Copy
83
Appliccation of Mod
when therre is a hypotthesized chain
n of events that
t
leads to acute toxicityy. Specificallly, as describbed in
Chapter 3,
3 nontesting approaches
a
co
ould be develloped for inittial or intermeediate events along a mechhanistic pathwaay. The prediictions could then be inpu
uts into modeels that predicct acute toxiccity on the baasis of
biologicall assay data on
o the intermeediate events. As a result, ppredictions oof acute toxiciity that incorpporate
biologicall data could be
b made for ch
hemicals for which
w
biologiical assays haave not yet beeen conductedd.
FIGURE 5-3
5 ToxPi mod
del for integrattion of acute-to
oxicity potentiial. In this simuulated examplee, data from noontesting approaaches (Chapterr 3) and assayss (Chapter 4) have been inteegrated into a crossend point ToxPi moddel for
pulmonary
y toxicity, neurrotoxicity, and death. The key
y (lower panell, inset) shows that evidence for the two D
Death
slices havee been given ex
xtra weight in determining
d
th
he overall integgrated ranking for acute toxiccity. For each cchemical profile,, the distance of
o a slice from the origin ind
dicates the relattive potency. T
The longer sllices indicate cchemicals that arre more potent than chemicalls that have sh
horter slices oor those deemed inactive (inddicated by abseence of
a given slicce). The upperr panel shows all
a profiles in rank
r
order. Thee lower panel ttranslates the pprofiles into a pplot of
the integraated ranks vs sccores, with 95%
% confidence in
ntervals extendding from the rred square reprresenting each chemical. The black-framed
b
prrofile is the ch
hemical that haas the highest ooverall acute-tooxicity potential (that is, rankk = 1).
The red-fraamed profile iss the reference chemical for high
confidencce of high toxiicity, and the green-framed pprofile

is the referrence chemical for high con
nfidence of low
w toxicity. N
Note that the coonfidence interrvals extendingg from
each refereence chemical define
d
the cateegory threshold
ds (vertical dashhed red and grreen lines).
84

For example, Chapter 3 discussed how chemical toxicity often results from nonspecific alterations in
cell function; thus, in vitro cytotoxicity is likely to be a strong indicator of in vivo toxicity. Conceivably, a
sequential model could be developed that combines a model that uses physicochemical data to predict
cytotoxicity (reviewed in Chapter 3) with a model that uses cytotoxicity data to predict LD50s (reviewed
in Chapter 4). In the future, one might envision using physicochemical information to predict bioactivity
in more specific biological assays and then using existing models that use bioactivity measurements to
predict in vivo acute toxicity.
Another area in which sequential modeling is common is in vitroin vivo extrapolation (IVIVE). In
particular, the results of in vitro assays constitute inputs into a reverse-dosimetry or reverse-toxicokinetic
approach to derive the external dose that will result in the internal serum concentration equivalent to the
bioactive concentration in the in vitro assay (Rotroff et al. 2010; Wetmore et al. 2012). Integration of
ADME specifically is discussed further below.
Sequential modeling is not restricted to single assay results. The US Environmental Protection
Agency (EPA) ToxCast program has generated in vitro high-throughput screening data on several assay
technologies that assess multiple pathways, genes, and responses over hundreds of end points (Judson et
al. 2010; Kavlock et al. 2012). The built-in redundancy in end points allows assays to be aggregated into a
pathway context, so that multiple assay results are combined into a summary measure of pathway activity
before a reverse toxicokinetic model is applied (Judson et al. 2011). Other researchers have attempted to
integrate assays anchored to pathways to arrive at a summary outcome, specifically for skin sensitization
(see, for example, Jaworska et al. 2013; Patlewicz et al. 2014; van der Veen et al. 2014).
Integrating Toxicokinetics to Determine Acute-Toxicity Potential
The recent investment in in vitro and high-throughput screening (HTS) strategies to inform chemical
toxicity testing has led to increased debate about relating the resulting data (potency values derived on the
basis of the nominal testing concentration ranges used in the wells of assay plates) to values that would be
informative in predicting in vivo human health hazard. Some assay designs do not consider in vivo toxicokinetic processes, which ultimately dictate the extent of chemical toxicity or potency in animals or humans. No matter how active or potent a chemical might be in some in vitro assays, if it is not absorbed
into the human body on exposure it will not be bioavailable to elicit any effect. Similarly, a chemical that
is cleared rapidly might not be present in the body long enough to elicit an effect. Chapters 3 and 4 describe how modeling of biological (toxicokinetic) processes can reduce the gaps observed between results
of in vitro assay and toxic response in humans or animals. Consideration of the potential effect of in vivo
toxicokinetic processes improves interpretation of the results of in vitro testing and the overall predictivetoxicology approach.
Development of toxicokinetic and dosimetric tools to inform extrapolationsamong species, dose
ranges, and experimental systems (for example, in vitro to in vivo)has been widespread over the last 30
years. Forward dosimetry with physiologically based pharmacokinetic modeling originally introduced in
the assessment of volatile organic chemicals (Andersen et al. 1987) provides a strategy that relies on
pharmacokinetic knowledge to relate a known external exposure to an internal blood or target-tissue dose.
Alternatively, reverse dosimetry is often used to relate a known internal dose (either from blood biomonitoring data or from an in vitro assay bioactivity concentration) to an external chemical dose (Tan et al.
2007; Lyons et al. 2008; Wetmore et al. 2012). The focus of much of the reverse-dosimetry work has
been on exposure scenarios (chronic, low-level, repeat exposures) of concern to EPA or other regulatory
bodies that have similar public-health protection mandates (Basketter et al. 2012; Wetmore in press).
Whereas it is useful to understand the dosimetric strategies described, it should be noted that not all tools
are directly applicable for DODs purpose in which the exposure will be acute (probably at one time) and
that acutely toxic chemicals might not require consideration of ADME to predict elicitation of debilitating
effects. The following discussion provides a brief summary of the different components of ADME, tools
Prepublication Copy
85
available to predict the components, and probable effect of ADME on future attempts by DOD to predict
acute toxicity.
Absorption or Bioavailability. Absorption via relevant routes of exposure (oral, dermal, and inhalation) and resulting bioavailability are important parameters for which there are predictive testing and
nontesting tools (see Chapters 3 and 4). A conservative assumption of 100% absorption would be the
most protective and should be assumed for chemicals for which available methods do not apply or do not
provide sufficient predictivity. Computer models that describe the biochemical behavior of uptake by skin
or gastrointestinal cells (McKone 1990; Jamei et al. 2009a; Rauma et al. 2013) can be applied in a sequential approach. However, the uptake models have been developed mostly for single-chemical exposure
and rarely for acute conditions. If 100% absorption was initially assumed for a chemical and predictive
tools indicate an adjustment away from that conservative default, the chemical might need to be downgraded or reassessed because lower absorption would indicate a lower potency or lower likelihood of
acute toxicity. Chemicals that have other toxicity alerts that place them in the high confidence of high
toxicity bin and identify them as readily absorbed should be noted because confirmation of high bioavailability will influence their ranking in that bin.
Distribution. Rapid accumulation throughout the body or in target tissues (such as skin, brain, and
lungs) could lead to a substantial shift in acute-toxicity potential of a chemical. Some organophosphates,
for example, are highly lipophilic and are readily distributed into fat and other tissues. Presence in the fat
and slow release from the site will delay or prolong acetylcholinesterase inhibition (Karalliedde et al.
2003) and could lead to greater toxicity than that of chemicals that have lower distribution but a similar
effect in an in vitro assay. Similarly, a chemical that is demonstrably neurotoxic in an in vitro assay and is
identified as crossing the bloodbrain barrier rapidly will likely be a more potent neurotoxicant than one
that has similar in vitro toxicity and does not cross the bloodbrain barrier. Blood transporters or lipoproteins can also shift the ability of a chemical to reach a toxicity target. Chemicals that bind to transporter
proteins might reach target sites more easily than chemicals that are distributed solely by diffusion. Tools
for predicting or assessing tissue partitioning and partition coefficients are available and can aid in predicting ADME behavior (Poulin and Haddad 2012, 2013). Distribution and metabolism (see below) are
probably the two most important components to explore in modulating acute-toxicity potential. Distribution will indicate the amount of a chemical that is available to reach an in vivo targeted site. Modeling
tools and simple in vitro experiments can estimate tissue or cell partitioning and thus provide a more accurate estimate of assay concentration that will generate a response.
Metabolism. Metabolism can lead to formation of a substantially more toxic metabolite, formation
of an equally toxic metabolite, or detoxification of the parent chemical. Examples of chemicals that can
be bioactivated into more toxic metabolites are the neurotoxic organophosphate insecticides, such as fenthion, parathion, diazinon, malathion, and chlorpyrifos, which are metabolized to potent oxon metabolites
(Eisler 2007). Thus, metabolism is one ADME property that could substantially shift the potency of a
chemical. Nontesting approaches for predicting metabolism are in various stages of development (see
Chapter 3) and might be used to elucidate metabolites that could contribute to a parent chemicals acutetoxicity potential. The best way to incorporate the potential effect of toxic metabolite formation will depend largely on the chemical mechanism in relation to others that have similar acute-toxicity potencies
and will need to be considered in relation to other toxicity information.
Excretion. The liver and kidney are the major organs responsible for excretion of chemicals from
the body (via bile and urine). Physicochemical properties can be used to predict chemical excretion rates
(Ghibellini et al. 2006; Shitara et al. 2006; Sharifi and Ghafourian 2014). Hydrophilic chemicals are more
rapidly excreted from the body than lipophilic chemicals. The other ADME properties are more likely
than excretion to have an important effect on chemical acute-toxicity potential. Reductions in excretion
are likely to be secondary to renal toxicity, an end point that is likely to be specifically assessed with other
testing and nontesting approaches.
86
Prepublication Copy

In Vitro to In Vivo Extrapolation Modeling to Inform Tissue Dosimetry andDosimetric Potential
Modeling approaches developed to inform dosimetry assessments use IVIVE. Measurements from
in vitro assays and predictions from nontesting approaches can provide various model inputs (such as rate
of absorption, metabolic activity, and tissue partitioning), which can be combined in a bottom-up approach to estimate an in vivo dose (Jamei et al. 2009b). The extrapolation of in vitro data typically assumes that metabolism of a parent chemical implies loss of potential for toxicity, which is not necessarily
the case, but it is valuable in providing an understanding of the dosimetrics or bioaccumulative potential
of a chemical.
Recently, a simplified IVIVE approach amenable to incorporation with HTS data was presented; it
predicted external doses that are required to achieve steady-state blood concentrations similar to ones that
elicit activity in in vitro HTS assays (Rotroff et al. 2010; Wetmore et al. 2012). The approach incorporated plasma-protein binding and hepatic metabolic and renal nonmetabolic clearance (key determinants
of chemical steady-state toxicokinetics) to estimate chemical steady-state behavior. The toxicokinetic
measurements showed not only significant cross-species correlation but strong correlation between in vivo and in vitro values for several chemicals (Wetmore et al. 2013, in press). Steady-state concentrations
are known to be more relevant for chronic and repeated exposure. Recent assessments by EPA demonstrated that in some cases steady-state concentration (Css) estimates were consistent with peak concentrations (Cmax); this relationship was not observed for a subset of chemicals that are rapidly or slowly cleared
(Wambaugh et al. in press). A steady-state toxicokinetic assumption, however, can be valid for acutetoxicity assessment.
Many of the tools developed to predict ADME or pharmacokinetic properties were developed by using pharmaceutical chemicals, which represent a smaller chemical space than the chemical domain of
concern to DOD. Tools to predict absorption, tissue partitioning, protein binding, and hepatic clearance
might perform well only in a narrowly defined logKow space. Chemicals of concern to DOD will span
multiple domains and attempts to categorize their pharmacokinetic behavior might be severely limited.
Ultimately, one should recognize that the most reactive acutely toxic chemicals that are of top concern to DOD are likely to be rapidly lethal well before some metabolic or excretory processes are initiated. However, for a large percentage of the chemical space of concern to DOD, application of some of the
physicochemical and assay tools to predict chemical toxicokinetics can probably be incorporated into the
decision-making process reasonably efficiently. Overall, integration of the tools in a tiered testing framework will aid in refining estimates of the toxicity of chemicals and enable an appropriate and streamlined
decision-making process.
APPROACHES TO CATEGORIZATION
As discussed in Chapter 2, the committees suggested prioritization strategy consists of a tiered approach to placing chemicals into three categories: high confidence of high toxicity, high confidence of
low toxicity, and inadequate data. As described earlier in this chapter, the process includes two key
steps: setting benchmarks that define the thresholds for what would be considered high and low toxicity
and assigning chemicals to categories, taking into account the confidence in the prediction.
Setting Benchmarks
The thresholds for high and low toxicity for a given toxicity end point can be defined in multiple
ways, including the following:
Using reference chemicals of high and low toxicity when the toxicity end point of interest has
been measured or predicted. This approach was illustrated in the simplified example in Box 5-1 and in
Figure 5-1 with LD50 as the toxicity end point.
Prepublication Copy
87
Using reference chemicals of high and low toxicity via clustering approaches. For example, the
nearest neighbors to the high toxicity reference chemicals would also be considered to have high toxicity.
Using pre-existing exposure-based thresholds if they exist for the toxicity end point of interest.
For example, the European Union (EU) and the Globally Harmonized System of Classification and Labeling of Chemicals (GHS) each have pre-existing categories for acute oral toxicity in terms of milligrams
per kilogram.
All the approaches require that the toxicity end point be a numerical value, such as an LD50 in milligrams
per kilogram, and they do not work for qualitative or binary outputs, such as active and inactive.
Categorization Decisions under Uncertainty
With respect to assigning chemicals to categories and taking into account confidence, by definition,
the assignments to the high and low toxicity categories need to have high confidence. It is therefore of
paramount importance for the categorization process to characterize uncertainty in the prediction of the
toxicity end point.
Uncertainty is an unavoidable aspect of knowledge discovery from large-scale heterogeneous data,
such as those discussed in Chapters 3 and 4. It has two main sources: the data themselves and the modeling of those data. With respect to data, even physicochemical properties cannot be measured with perfect
precision. Moreover, such data are usually obtained from databases, which can contain errors that are due
to data entry or other processes related to creating and populating the databases (Fourches et al. 2010).
For biological assays, uncertainties will vary considerably with the type of assay. Although automation of
some assays has greatly improved their reproducibility, there is still variation among batches from a given
assay and possibly even among chemicals in a given batch. Uncertainty arising from the modeling or
analysis approach (different aspects of data handling and data analysis) varies with the model and among
parameter values in a given model.
There is a rich literature on characterizing uncertainty in experimental and computational sciences
that need not be recapitulated here. Suffice it to say that the existing approaches span a wide range that is
based on both frequentist and Bayesian statistical principles and use analytical methods (for example,
classical confidence intervals) and sampling or simulation-based methods (such as bootstrap and Monte
Carlo). In some cases, specific guidance is available on characterizing uncertainty for a particular application, such as QSAR modeling, gene-expression analyses, and pharmacokinetic modeling (IPCS 2008).
For example, Chapter 3 discussed the availability of Organisation for Economic Co-operation and Development (OECD) and Registration, Evaluation, Authorization and Restriction of Chemicals (REACH)
guidance on evaluating the confidence in (Q)SAR models. Although internal and external cross-validation
make up the current standard approach to characterizing uncertainty in QSAR predictions, alternative approaches are being pursued. For example, Gramacy and Pantaleo (2010) used a Markov Chain Monte Carlo
sampling for the Bayesian Lasso model to assess prediction uncertainties in QSAR analyses. More discussion of other aspects of uncertainty in QSAR predictions can be found in Sahlin (2013). Parametric methods
can also be used sometimes, but these might be less appealing for complex data from multiple sources because distributional assumptions are likely to be violated.
An additional useful tool for assessing confidence is sensitivity analysis, which can tell how the uncertainty in the toxicity prediction of a model or system can be apportioned to different sources of uncertainty in its inputs or, equivalently, how much each input is contributing to the uncertainty. For example,
sensitivity analyses might be performed for several purposes:
To test the robustness of results by random data perturbation.
To decompose the prediction error by relating input and output variables in a model, which can
help to understand the sources of variation in the model. Such analyses can identify inputs that are key
88
Prepublication Copy

contributors to uncertainty and thereby focus research or data generation on aspects that could improve or
refine the model.
To perform model selection through internal cross-validation or external validation. Internal
cross-validation can also be used to quantify the degree of fitting or overfitting of a model. External validation (meta-analysis and related data streams) is preferred as an independent assessment of a models
predictive ability.
Finally, it should be noted that these approaches to assessment of uncertainty and sensitivity might
be applied not only to toxicity end point predictions but to the categorization decisions themselves. Application to the categorization decision is illustrated in Box 5-2, where uncertainty is analyzed by using
the irreproducible discovery rate.
FINDINGS
Finding: The committees recommended prioritization strategy will require integration at various
stages, and there are many approaches to integration, including formal statistical methods (such as metaanalysis), less formal recombination methods (such as ToxPi), methods that pool datasets, methods that
use datasets hierarchically, and methods that link models sequentially.
Finding: There are several possible approaches to placing chemicals in categories of high toxicity, low toxicity, and inadequate data, including quantitative thresholds based on reference chemicals (such as sarin), generic thresholds based on external criteria (such as those of the EU and the GHS),
and clustering based on reference chemicals. The stability of and confidence in the results of any categorization will be enhanced if the uncertainties are properly quantified.
Finding: Multiple levels of complexity require integration, and the committee has distinguished
between integration of predictions among end points (crossend point integration) and integration of different model predictions or data streams that inform a single end point (withinend point integration).
Finding: Given the many types of end points that are relevant to acute, debilitating toxicity and
the need to place chemicals into categories of high toxicity, low toxicity, and inadequate data, the
simplest approach to crossend point integration would be to summarize the categorization results for
each end point in a scorecard.
Finding: Some of the key policy decisions that DOD will need to make to use the committees
recommended prioritization strategy are (1) the kinds of responses that would be considered appropriate
end points (for example, neurotoxicity, seizures, or cholinesterase inhibition), (2) determination of high
and low toxicity thresholds for each end point of interest, (3) the degree of confidence required to conclude high confidence, and (4) decision rules related to a determination of a summary conclusion of high
or low toxicity on the basis of multiple end points.
Finding: Toxicokinetic and ADME behavior can influence the prediction of a chemicals acute
toxicity potential and resulting categorization, and this emphasizes the need to include such considerations into the tiered prioritization strategy.
REFERENCES
Andersen, M.E., H.J. Clewell III, M.L. Gargas, F.A. Smith, and R.H. Reitz. 1987. Physiologically based pharmacokinetics and the risk assessment process for methylene chloride. Toxicol. Appl. Pharmacol. 87(2):185-205.
Basketter, D.A., H. Clewell, I. Kimber, A. Rossi, B. Blaauboer, R. Burrier, M. Daneshian, C. Eskes, A. Goldberg, N.
Hasiwa, S. Hoffmann, J. Jaworska, T.B. Knudsen, R. Landsiedel, M. Leist, P. Locke, G. Maxwell, J. McKim,
E.A. McVey, G. Oudraogo, G. Patlewicz, O. Pelkonen, E. Roggen, C. Rovida, I. Ruhdel, M. Schwarz, A.
Schepky, G. Schoeters, N. Skinner, K. Trentz, M. Turner, P. Vanparys, J. Yager, J. Zurlo, and T. Hartung.
2012. A roadmap for the development of alternative (non-animal) methods for systemic toxicity testing - t4
report. ALTEX 29(1):3-91.
Prepublication Copy
89
Bauer-Mehren, A., E.M. van Mullingen, P. Avillach, M.D.C. Carrascosa, R. Garcia-Serna, J. Piero, B. Singh, P.
Lopes, J.L. Oliveira, G. Diallo, E.A. Helgee, S. Boyer, J. Mestres, F. Sanz, J.A. Kors, and L.I. Furlong. 2012.
Automatic filtering and substantiation of drug safety signals. PLoS Comput. Biol. 8(4):e1002457.
Borenstein, M., L.V. Hedges, J.P.T. Higgins, and H.R. Rothstein. 2009. Introduction to Meta-Analysis. Chichester,
U.K: Wiley.
Crowther, M., W. Lim, and M.A. Crowther. 2010. Systematic review and meta-analysis methodology. Blood
116(17):3140-3146.
Diaza, R.G., S. Manganelli, A. Esposito, A. Roncaglioni, A. Manganaro, and E. Benfenati. 2015. Comparison of in
silico tools for evaluating rat oral acute toxicity. SAR QSAR Environ Res. 26(1):1-27.
Eisler, R. 2007. Eisler's Encyclopedia of Environmentally Hazardous Priority Chemicals. Amsterdam: Elsevier.
EPA (US Environmental Protection Agency). 2014. Meeting on Scientific Issues Associated with Integrated Endocrine Bioactivity and Exposure based Prioritization and Screening, December 2-5, 2014. FIFRA Scientific
Advisory Panel, US Environmental Protection Agency [online]. Available: http://www.epa.gov/scipoly/sap/
meetings/2014/december/120214transcript.pdf [accessed March 26, 2015].
Fourches, D., E. Muratov, and A. Tropsha. 2010. Trust, but verify: On the importance of chemical structure curation
in cheminformatics and QSAR modeling research. J. Chem. Inf. Model. 50(7):1189-1204.
Ghibellini, G., E.M. Leslie, and K.L. Brouwer. 2006. Methods to evaluate biliary excretion of drugs in humans: An
updated review. Mol. Pharm. 3(3):198-211.
Gramacy, R.B., E. Pantaleo. 2010. Shrinkage regression for multivariate inference with missing data, and an application to portfolio balancing. Bayesian Anal. 5(2):237-262.
Hedges, L.V., and I. Olkin. 1985. Statistical Methods for Meta-Analysis. New York: Academic Press.
Higgins, J.P.T., and S. Green, eds. 2011. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0
[updated March 2011]. The Cochrane Collaboration [online]. Available: http://handbook.cochrane.org/ [accessed March 24, 2015].
IPCS (International Programme on Chemical Safety). 2008. Uncertainty and Data Quality in Exposure Assessment.
World Health Organization [online]. Available: http://www.who.int/ipcs/publications/methods/harmonization/
exposure_assessment.pdf?ua=1 [accessed March 24, 2015].
Jamei, M., D. Turner, J. Yang, S. Neuhoff, S. Polak, A. Rostami-Hodjegan, and G. Tucker. 2009a. Population-based
mechanistic prediction of oral drug absorption. AAPS J. 11(2):225-237.
Jamei, M., S. Marciniak, K. Feng, A. Barnett, G. Tucker, and A. Rostami-Hodjegan. 2009b. The Simcyp population-based ADME simulator. Exp. Opin. Drug Metab. Toxicol. 5(2):211-223.
Jaworska, J., Y. Dancik, P. Kern, F. Gerberick, and A. Natsch. 2013. Bayesian integrated testing strategy to assess
skin sensitization potency: From theory to practice. J. Appl. Toxicol. 33(11):1353-1364.
Judson, R.S., K.A. Houck, R.J. Kavlock, T.B. Knudsen, M.T. Martin, H.M. Mortensen, D.M. Reif, D.M. Rotroff, I.
Judson, R.S., R.J. Kavlock, R.W. Setzer, E.A. Cohen Hubal, M.T. Martin, T.B. Knudsen, K.A. Houck, R.S. Thomas, B.A. Wetmore, and D.J. Dix. 2011. Estimating toxicity-related biological pathway altering doses for highthroughput chemical risk assessment. Chem. Res. Toxicol. 24(4):451-462.
Karalliedde, L.D., P. Edwards, and T.C. Marrs. 2003. Variables influencing the toxic response to organophosphates
in humans. Food Chem. Toxicol. 41(1):1-13.
Kavlock, R., K. Chandler, K. Houck, S. Hunter, R. Judson, N. Kleinstreuer, T. Knudsen, M. Martin, S. Padilla, D.
Reif, A. Richard, D. Rotroff, N. Sipes, and D. Dix. 2012. Update on EPA's ToxCast program: Providing high
Lee, S., K. Park, H.S. Ahn, and D. Kim. 2010. Importance of structural information in predicting human acute toxicity from in vitro cytotoxicity data. Toxicol. Appl. Pharmacol. 246(1-2):38-48.
Li, Q., J.B. Brown, H. Huang, and P.J. Bickel. 2011. Measuring reproducibility of high-throughput experiments.
Ann. Appl. Stat. 5(3):1752-1779.
Linkov, I., O. Massey, J. Keisler, I. Rusyn, and T. Hartung. 2015. From weight of evidence to quantitative data
integration using multicriteria decision analysis and Bayesian methods. ALTEX 321(1):3-8.
Lounkine, E., M.J. Keiser, S. Whitebread, D. Mikhailov, J. Hamon, J.L. Jenkins, P. Lavan, E. Weber, A.K. Doak, S.
Ct, B.K. Shoichet, and L. Urban. 2012. Large-scale prediction and testing of drug activity on side-effect targets. Nature 486(7403):361-367.
Lyons, M., R.S. Yang, A.N. Mayeno, and B. Reisfeld. 2008. Computational toxicology of chloroform: Reverse dosimetry using Bayesian inference, Markov Chain Monte Carlo simulation, and human biomonitoring data. Environ. Health Perspect. 116(8):1040-1046.
90
Prepublication Copy

McKone, T.E. 1990. Dermal uptake of organic chemicals from a soil matrix. Risk Anal. 10(3):407-419.
NRC (National Research Council). 2014. A Framework to Guide Selection of Chemical Alternatives. Washington,
DC: National Academies Press.
Patlewicz, G., C. Kuseva, A. Kesova, I. Popova, T. Zhechev, T. Pavlov, D.W. Roberts, and O. Mekenyan. 2014.
Towards AOP application--implementation of an integrated approach to testing and assessment (IATA) into a
pipeline tool for skin sensitization. Regul. Toxicol. Pharmacol. 69(3):529-545.
Pavan, M., and A. Worth. 2008. A Set of Case Studies to Illustrate the Applicability of DART (Decision Analysis by
Ranking Techniques) in the Ranking of Chemicals. EUR 23481. Luxembourg: European Communities
[online]. Available: https://eurl-ecvam.jrc.ec.europa.eu/laboratories-research/predictive_toxicology/doc/EUR_
23481_EN.pdf
Poulin, P., and S. Haddad. 2012. Advancing prediction of tissue distribution and volume of distribution of highly
lipophilic compounds from a simplified tissue-composition-based model as a mechanistic animal alternative
method. J. Pharm. Sci. 101(6):2250-2261.
Poulin, P., and S. Haddad. 2013. Hepatocyte composition-based model as a mechanistic tool for predicting the cell
suspension: Aqueous phase partition coefficient of drugs in in vitro metabolic studies. J. Pharm. Sci.
102(8):2806-2818.
Rauma, M., A. Boman, and G. Johanson. 2013. Predicting the absorption of chemical vapours. Adv. Drug Deliv.
Rev. 65(2):306-314.
Reif, D.M., M.T. Martin, S.W. Tan, K.A. Houck, R.S. Judson, A.M. Richard, T.B. Knudsen, D.J. Dix, and R.J.
Kavlock. 2010. Endocrine profiling and prioritization of environmental chemicals using ToxCast data. Environ. Health Perspect. 118(12):1714-1720.
Rotroff, D.M., B.A. Wetmore, D.J. Dix, S.S. Ferguson, H.J. Clewell, K.A. Houck, E.L. Lecluyse, M.E. Andersen,
R.S. Judson, C.M. Smith, M.A. Sochaski, R.J. Kavlock, F. Boellmann, M.T. Martin, D.M. Reif, J.F. Wambaugh, and R.S. Thomas. 2010. Incorporating human dosimetry and exposure into high-throughput in vitro
toxicity screening. Toxicol. Sci. 117(2):348-358.
Sahlin, U. 2013. Uncertainty in QSAR predictions. ATLA 41(1):111-125.
Sedykh, A., H. Zhu, H. Tang, L. Zhang, A. Richard, I. Rusyn, and A. Tropsha. 2011. Use of in vitro HTS-derived
concentration-response data as biological descriptors improves the accuracy of QSAR models of in vivo toxicity. Environ. Health Perspect. 119 (3):364-370.
Sharifi, M., and T. Ghafourian. 2014. Estimation of biliary excretion of foreign compounds using properties of molecular structure. AAPS J. 16(1):65-78.
Shitara, Y., T. Horie, and Y. Sugiyama. 2006. Transporters as a determinant of drug clearance and tissue distribution. Eur. J. Pharm. Sci. 27(5):425-446.
Tan, Y.M., K.H. Liao, and H.J. Clewell III. 2007. Reverse dosimetry: Interpreting trihalomethanes biomonitoring
data using physiologically based pharmacokinetic modeling. J. Expo. Sci. Environ. Epidemiol. 17(7):591-603.
van der Veen, J.W., E. Rorije, R. Emter, A. Natsch, H. van Loveren, and J. Ezendam. 2014. Evaluating the performance of integrated approaches for hazard identification of skin sensitizing chemicals. Regul. Toxicol. Pharmacol. 69(3):371-379.
Wambaugh, J.F., B.A. Wetmore, R. Pearce, C. Strope, R. Goldsmith, J.P. Sluka, A. Sedykh, A. Tropsha, S. Bosgra,
I. Shah, R. Judson, R.S. Thomas, and R.W. Setzer. In press. Toxicokinetic triage for environmental chemicals.
Toxicol. Sci.
Wetmore, B.A. In press. Quantitative in vitro-to-in vivo extrapolation in a high-throughput environment. Toxicology.
Wetmore, B.A., J.F. Wambaugh, S.S. Ferguson, M.A. Sochaski, D.M. Rotroff, K. Freeman, H.J. Clewell III, D.J.
Dix, M.E. Andersen, K.A. Houck, B. Allen, R.S. Judson, R. Singh, R.J. Kavlock, A.M. Richard, and R.S.
Thomas. 2012. Integration of dosimetry, exposure, and high-throughput screening data in chemical toxicity
assessment. Toxicol. Sci. 125(1):157-174.
Wetmore, B.A., J.F. Wambaugh, S.S. Ferguson, L.Li, H.J. Clewell, III, R.S. Judson, K. Freeman, W. Bao, M.A.
Sochaski, T.M. Chu, M.B. Black, E. Healy, B. Allen, M.E. Andersen, R.D. Wolfinger, and R.S. Thomas.
2013. The relative impact of incorporating pharmacokinetics in predicting in vivo hazard and mode-of-action
from high-throughput in vitro toxicity assays. Toxicol. Sci. 132(2):327-346.
Wetmore, B.A., B. Allen, H.J. Clewell, III, T. Parker, J.F. Wambaugh, L.M. Almond, M.A. Sochaski, and R.S.
Thomas. 2014. Incorporating population variability and susceptible subpopulations into dosimetry for highthroughput toxicity testing. Toxicol. Sci. 142(1):210-224.
Wilson, A., D.M. Reif, and B.J. Reich. 2014. Hierarchical dose-response modeling for high-throughput toxicity
screening of environmental chemicals. Biometrics 70(1):237-246.
Zhang, S. 2011. Computer-aided drug discovery and development. Meth. Mol. Bio. (716):23-38.
Prepublication Copy
91
Zhu, H. L. Ye, A. Richard, A. Golbraikh, F.A. Wright, I. Rusyn, and A. Tropsha. 2009. A novel two-step hierarchical quantitative structure-activity relationship modeling work flow for predicting acute toxicity of chemicals in rodents. Environ. Health Perspect. 117:(8):1257-1264.
92
Prepublication Copy
6
Lessons Learned and Next Steps
The objective of the committees conceptual framework as presented in Chapter 2 is to predict the
potential of a chemical to cause acute toxicity to organ systems that could result in debilitating or lethal
effects.1 In developing its framework, the committee considered how to characterize the inherent toxicity
of a chemical, evaluate metabolic and pharmacokinetic attributes that can modify chemical toxicity, and
integrate information over different domains. As discussed in Chapter 3, the conceptual framework includes models that use physicochemical and biological data to make predictions about potential acute toxicity. The predictive models can be qualitative (such as structural alerts) or quantitative (such as quantitative structureactivity relationship [QSAR] models), and the resulting outputs themselves might be
qualitative (for example, a toxic or nontoxic determination) or quantitative (for example, a potency estimate). Medium-throughput and high-throughput assays will also be needed to predict acute mammalian
toxicity as reviewed in Chapter 4. The toxicity predictions ultimately will depend on the collection and
integration of the physicochemical and biological data that might be indicators of potential acute toxicity
as discussed in Chapter 5.
One primary goal of the committees conceptual framework is to develop data sufficient to categorize chemicals on the basis of their predicted acute toxicity. The committee considered existing toxicitybased chemical classification schemes developed for industrial chemicals, agrochemicals, biocides, and
pharmaceuticals that often use lethality data to estimate toxicity, such as the dose required to kill 50% of a
population of test animals (LD50) (Seidle et al. 2010). Developing LD50s was considered a useful benchmarking approach for predicting acute toxicity of chemicals of interest to the Department of Defense
(DOD) because it allows DOD to exclude low-toxicity chemicals and to focus its resources on more toxic
chemicals of concern. The committee considered the need to develop mechanistically based assays and to
use well-characterized chemicals as positive controls to improve the predictive validity of LD50s and to
identify potential organ toxicities.
In the present chapter, the committee briefly discusses some current programs that are evaluating or
developing modern testing strategies, reviews its suggested tiered testing strategy, and highlights important lessons learned from current testing programs that should be considered by DOD in developing its
future testing strategy. The chapter also provides the committees overall conclusions and identifies several steps that DOD could take in the short term to medium term (310 years) to implement a program
that uses modern approaches to identify chemicals that have the potential to induce life-threatening acute
toxicity in deployed personnel. 2
MODERN APPROACHES FOR THE ASSESSMENT OF ACUTE CHEMICAL TOXICITY
The committee explored whether DOD could adopt a modern testing strategy for the prediction of
1
As defined in Chapters 1 and 2, organ systems included the cardiovascular, respiratory, hepatic, renal, skeletomuscular, immune, and nervous systems, including special senses (vision and hearing).
2
The committee recognizes that a more detailed research plan is needed and that development of such a plan is
noted in the statement of task as a potential Phase 2 of this project.
Prepublication Copy
93
acute toxicity.3 In particular, the committees statement of task required a focus on the assessment of existing high-throughput screening (HTS) methods to identify acutely (and likely highly) toxic chemicals
with greater predictivity. The committee considered whether several projects were relevant for DODs
purposes. For example, it examined the European Centre for the Validation of Alternative Methods
(ECVAM) ACuteTox project4 whose stated aim is to develop a strategy to replace all in vivo tests of
acute oral toxicity. The ACuteTox effort considers in vitro methods that address specific mechanisms of
action relevant to acute systemic toxicity (such as assays for neurotoxicity) and includes such computational methods as QSAR modeling and physiologically based biokinetic modeling.
Initially, the ACuteTox project selected and tested 97 reference chemicals with six basal cytotoxicity
assays and compared the results with published human and animal in vivo data (Clothier et al. 2008;
Sjstrm et al. 2008). Later, 57 reference chemicals were tested in a number of functional tests that covered
absorption, distribution, metabolism, excretion, and specific organ and system toxicity, such as hematotoxicity, neurotoxicity, nephrotoxicity, and hepatotoxicity (Kinsner-Ovaskainen et al. 2009). Standardized experimental design and data acquisition were used in the second phase of ACuteTox program (KoppSchneider et al. 2013). Concentrationresponse data were routinely collected (for example, IC20, IC50, EC20,
or EC50 values) and served as the statistical basis of the development of testing strategies (KinsnerOvaskainen et al. 2013; Prieto et al. 2013b). The oral acute-toxicity category was predicted using a chemicals physicochemical properties, in silico modeling results, and values (such as IC50) obtained from the in
vitro studies (Kopp-Schneider et al. 2013). In general, the predictive validity seen in the efforts has been
moderate to low. In addition, there has been little effort to assess the predictive validity for highly toxic
chemicals that would be of concern to DOD.
Examples of large-scale US projects include the Toxicology Testing in the 21st Century (Tox21) partnership and the US Environmental Protection Agency (EPA) ToxCast program. The ToxCast program uses
a large suite of in vitro biochemical (cell-free) and cell-based assays to evaluate chemicals and to analyze
their bioactivity profiles computationally (Kavlock et al. 2012; Judson et al. 2014; Kleinstreuer et al. 2014).
Phase I of the ToxCast program included about 300 conventional pesticide active ingredients that were tested in a battery of cell-free and cell-based assays to evaluate the ability of the assays to predict potential human toxicity (Judson et al. 2010). In Phase II, the chemical space was broadened to include chemicals that
are used in consumer products and industrial processes and unmarketed drugs that were donated by pharmaceutical companies (Sipes et al. 2013). However, few ToxCast assays were designed specifically to assess
acute toxicity. Furthermore, there have been few efforts to use HTS approaches to evaluate acute toxicity.
Regardless, the ToxCast program has identified a number of important technical issues that could be considered in the design of a program relevant for DOD (Tice et al. 2013).
The committee also explored modern toxicity-testing strategies that are under development in the
pharmaceutical industry. For example, the pharmaceutical industry is exploring ways to predict druginduced liver injury (DILI). DILI is a low-incidence but important idiosyncratic cause of drug toxicity and
a major reason for attrition during drug development or withdrawal from the market (Stephens et al.
2014). Traditional toxicity-testing strategies do not predict DILI reliably in patients: fewer than 55% and
25% of DILI drugs are predicted on the basis of the regulatory animal-toxicity studies and simple in vitro
tests, respectively (Olson et al. 2000; Xu et al. 2004). Box 6-1 describes the recent development of predictive models that can be used in preclinical studies to detect DILI risk in humans.
3
Modern approaches are ones that do not rely primarily on traditional toxicology testing and include computational and in vitro or nontraditional in vivo assays.
4
See http://www.acutetox.eu/. The ACuteTox consortium was initially funded with 8 million by the European
Commissions 6th Framework Programme and included 35 academic, industrial, and government research institutes
in 13 European countries (Clemedson 2008). The project was divided into 10 work packages that included selection
of reference chemicals, generation of animal and human in vivo databases, development of an Internet-based database for central management of all project data, adaptation of promising in vitro methods to robotic screening platforms, statistical analysis (identification of outliers and design of the preliminary algorithm or prediction model),
development of in vitro assays for neurotoxicity, and construction and optimization of the in vitro testing strategy.
94
Prepublication Copy

The committee was unable to find a robust modern testing program that DOD could readily adopt
for its purposes. Lessons learned from the ToxCast, ACuteToX, and industry programs do, however, provide a great deal of guidance for DOD in designing a system that uses HTS approaches and predictive
models to interpret the performance of cell-free and cell-based assays in predicting acute toxicity.
IMPLEMENTATION OF THE COMMITTEES CONCEPTUAL MODEL
FOR ASSESSMENT OF ACUTE CHEMICAL TOXICITY
As discussed in Chapter 2, the committee developed a conceptual framework that links chemical
structure, physicochemical properties, biochemical properties, and biological activity to acute toxicity.
Implementation of the conceptual framework will require DOD to support development of a suite of databases, assays, models, and tools that are based on in vitro, nonmammalian in vivo, and in silico approaches for predicting acute toxicity. As presented in Chapter 2, these databases, assays, models, and tools
would be used as part of a tiered prioritization strategy to predict acute toxicity (see Figure 6-1). Relatively easily obtained nontesting data (such as existing toxicity data or physicochemical properties; see
Chapter 3 for more details) could allow initial evaluation of a large number of chemicals in the initial tiers. Higher testing tiers incorporate a variety of assays (in vitro or nontraditional in vivo) whose biological complexity increases as a chemical moves from one tier to the next. At each tier, a decision is made as
to whether further assessment of toxicity is needed. The ultimate goal of the effort is to sort chemicals
into categories that would indicate the predicted acute-toxicity hazard (for example, low toxicity, high
toxicity, or uncertain toxicity because of inadequate data) and help to prioritize chemicals for follow-up
evaluation. DOD can also apply its expertise to use other factors that were not considered by the committee (such as chemical availability, ease of chemical synthesis, and weaponizability) to exclude chemicals
from further testing. One goal of the overall approach is to reduce the number of chemicals that progress
through each tier in an efficient and cost-effective manner.
BOX 6-1 An Example of an Integrated Testing Strategy for Predicting Drug-Induced Liver Injury
To facilitate identification of DILI drugs early in the drug-development process, Chen et al. (2014a)
developed a tool that combines exposure and physicochemical data and a panel of in vitro highcontent screening (HCS) assays to predict DILI. Implementation of the rule-of-two model (RO2) that
combines high exposure (100 mg/day) and high lipophilicity (logP 3) resulted in high specificity
(95%) but low sensitivity (27%); that equates to a high false-negative rate. The HCS panel alone,
which measured eight cellular end points (including apoptosis, cell loss, DNA damage, DNA fragmentation, and mitochondrial potential), was marginally more sensitive (39%). Integration of the RO2 model with the HCS assay panel increased the sensitivity to 55% (specificity remained at 95%).
Thoughtful consideration of the data sources or models to be integrated and of the compatibility of
the data streams from these sources is required to increase the likelihood of success of the prediction
strategy. Reported successes (Rusyn et al. 2012; Chen et al. 2014a) typically resulted from integration
of models that were based on different data sources that captured a greater diversity of information. In
Chen et al. (2014a), only six of the 27 positives were predicted in both the RO2 and HCS strategies;
this indicates the complementarity of the two approaches. Incorporation of mechanistic information into
DILI predictions can improve the model performance (Chen et al. 2014b).
Published hybrid approaches to integrating chemical descriptors with in vitro data have had little
success in improving predictions obtained on the basis of in vitro data alone (Low et al. 2011; Zhu et
al. 2014). Those approaches directly pool data from disparate sources for modeling purposes. Incorporation of different modeling strategies that retain more information from heterogeneous data structures in hybrid integrations would likely maximize the data available for assessment and result in improved predictivity. Similarly, use of adapters might facilitate data availability and integrity (Chen et al.
2014b).
Prepublication Copy
95
Tier0
Initial
Character
ization
(Chapter3)
Inadequatedata
+
Nontesting
approaches
(Chapter3)
Hightoxicity
Lowtoxicity
Lowthreat
duetofactors
unrelatedto
toxicity
Tier1
Integration
andDecision
Making
(Chapter5)
Inadequatedata
+
Biological
assaybased
approaches
(Chapter4)
Hightoxicity
Lowtoxicity
Lowthreat
duetofactors
unrelatedto
toxicity
Tier2
Integration
andDecision
Making
(Chapter5)
Hightoxicity
Lowtoxicity
Inadequatedata
Tier3
Lowthreat
duetofactors
toxicity
FIGURE 6-1 Prioritization strategy based on a tiered approach for using predictive-toxicology models and tools to
evaluate agents for acute toxicity.
DEVELOPMENT OF A MODERN TIERED APPROACH FOR

PREDICTING ACUTE TOXICITY: INITIAL STEPS
There are some initial steps that DOD could take to implement the committees tiered testing strategy. This section describes in greater detail some approaches that might be useful for DOD to pursue in
implementing modern testing approaches for predicting acute toxicity.
Computational Approaches
Chapter 3 discusses a variety of nontesting approaches that can be used to predict acute toxicity.
Physicochemical data can be used to predict physical hazards or reactivity, such pharmacokinetic characteristics as metabolism and distribution, and likely routes of exposure. Those data can help to group
chemicals by using chemical descriptors of various physicochemical or topological properties and to identify the biologically relevant chemical space (Lipinski and Hopkins 2004). Physicochemical data can also
help to fill data gaps and support read-across approaches that apply data on a particular property or effect
96
Prepublication Copy
Integration and
a Decision--Making for P
Predictive Toxxicology
of a tested
d chemical to a similar unttested chemiccal. A variety of in silico m
methods have been developped to
predict th
he molecular sites where metabolism could
c
occur and the typees of metabollites that couuld be
formed.5 Several
S
availaable QSAR models
m
use sttructural propperties or phyysicochemicall properties too predict acutee oral LD50s. Few such mo
odels are avaiilable for preedicting inhalation LC50s, and the comm
mittee
was unablle to identify models for prredicting derm
mal LD50s.
A feew QSAR mo
odels predict neurotoxicity
n
and cytotoxiccity but not oother end poinnts that are rellevant
for acute, debilitating toxicity.
t
Box 6-2 describess a QSAR moodel for the prrediction of accetylcholinesterase
(AChE) in
nhibition. Quantitative stru
ucturetoxicitty relationshipp models havve also been ddeveloped to eevaluate the role of lipophillicity, polarity
y, molecular geometry,
g
andd quantum chhemical descrriptors for moolecular orbitall energy in thee toxicity of organophosph
o
hate insecticiddes (Can 2014). Inhibitionn of AChE and other cholineesterases is an
n important sttep in the toxiicity of nervee-gas agents (ssuch as VX aand soman) thhat are
highly pottent organoph
hosphates; thu
us, the compu
utational moddels that have been developped could be uuseful
in a DOD testing strateegy.
A laarge number of chemicalss have reportted LD50 dataa, which covver a large poortion of cheemical
space. Altthough LD50 data are avaiilable on man
ny chemicals, they are not informative about a chem
mical's
mechanism
m of action. Such knowleedge is imporrtant becausee acute toxiciity might invvolve multiplee biochemical mechanisms; this highligh
hts the need fo
or improved m
mechanistic iinsights aboutt structuretooxicity
relationsh
hips. Efforts to
o derive acute-toxicity QS
SAR models tthat have highh predictive aaccuracy havee fallen short because
b
of meechanistic com
mplexities.
BOX
B
6-2 The Use of Comp
putational App
proaches for E
Evaluating
Chemica
al-Induced Inh
hibition of Ace
etylcholineste
erase
Recepto
or-specific sco
oring function
ns have been developed for predictin
ng binding afffinities of hum
man
acetylch
holinesterase (AChE) inhib
bitors. A study
y performed b
by Guo et al. (2004) illustrates the gen
neral
approac
ch used to develop predictive (Q)SAR models.
m
In thi s case, 69 ch
hemicals with
h IC50 data me
easured witth a human AChE
A
assay were
w
selected for training a
and testing of the scoring ffunction. The IC50
of the 69 chemicals ranged from 0.33 to 30,00
00 nM. Dockiing calculatio
ons were carrried out with p
published software (the Gold program
m). A weighte
ed sum of ele
ectrostatic an
nd van der W
Waals interacttions
between
n ligands and
d the receptorr residues wa
as calculated . Guo et al. e
examined the
e correlation of a
calculate
ed activity (pIC50) with exp
perimentally derived
d
IC50 data. The lefft figure show
ws the correla
ation
for a 53--ligand trainin
ng set (R2 = 0.89).
0
The right figure show
ws the resultss of using a novel 16-chem
mical
test set (R2 = 0.69).
Metabo
olites could alsso be identifieed through add
ditional experim
mentation, for example, by uusing chemicall analyses of bio
ological test sy
ystems or by co
omparing resullts obtained froom cell system
ms that use diffferent levels off metabolic comp
petence.
Prepubliccation Copy
97
High-Throughput Screens
How the Tox21 and ToxCast data have been used to predict mammalian toxicity was demonstrated
recently by the National Toxicology Program Interagency Center for the Evaluation of Alternative Toxicological Methods (NICEATM), which evaluated whether HTS data could predict the results of an in vivo uterotrophic assay that screens for estrogen activity in rodents (see Box 6-3). The project is part of a
larger NICEATM and EPA effort to develop a robust in vitro screen for chemicals with estrogenic or androgenic activity. The committee recognizes that estrogenic and androgenic end points are not relevant
for the assessment of acute toxicities of concern to DOD, but the NICEATM and EPA efforts provide
important insights. The committee found the targeted toxicity-prediction effort to be relatively successful
(see Box 6-3); this finding is promising for the DOD goal of acute-toxicity prediction. However, estrogenic activity is, in principle, an easier target for prediction because it depends on specific geneexpression patterns for which sensitive and specific HTS assays can be designed. Nevertheless, the effort
identified several needed features that can help to inform future DOD efforts, including the following:
Performance standards for new assays that consider validated test methods, reference chemicals,
and standards for assay accuracy and reliability (Wind and Stokes 2010; Stokes et al. 2012).
Suitable in vivo end points and study results that can be used to assess HTS assay performance
(Rotroff et al. 2013); that is, there is a need for phenotypic anchoring of the in vitro results.
Reference chemicals that have known biological activity (such as ER in the cited NICEATM
project) for evaluating the sensitivity, specificity, and predictivity of assays to identify agonists and antagonists for a biological target of interest (Huang et al. 2014).
Appropriate data-integration models that can pool information from multiple assays (Rotroff et al.
2013); model performance can be evaluated by calculating sensitivity, specificity, and balanced accuracy
for a specific set of criteria across chemical space (Rotroff et al. 2013; Cox et al. 2014).
Recognition that misclassification of chemicals might occur (Rotroff et al. 2013; Cox et al. 2014).
Possible explanations for misclassification of a chemical include a failure to test it at a high enough concentration to exhibit a response in ToxCast assays, inadequate metabolic competence of the test system,
and species, tissue, or cell-type differences in response between the ToxCast assay and in vivo models
that are used to assess HTS assay performance (Rotroff et al. 2013). Although not addressed directly by
the ToxCast program, an additional source of misclassification of a chemical could be the variability in
the in vivo data that are used to assess the ability of an assay to predict acute toxicity.
The example shown in Box 6-3 illustrates a current trend in the application of HTS testing as a replacement for some animal-based assays. In particular, the use of HTS assay data as part of an endocrine
screening program has received some traction in the scientific community (Dix et al. 2007; Thomas et al.
2012; Rotroff et al. 2013; Thomas et al. 2013; Cox et al. 2014; Becker et al. 2015). In some cases (such as
chemical interactions with the estrogen or androgen receptor system), adoption of HTS assays in the context of a tiered testing approach as a replacement for in vivo assays appears likely (van der Burg et al. in
press).
The predictiveness of many HTS assays, however, remains low (< 50%) (Thomas et al. 2012; Cox et
al. 2014; Patlewicz et al. 2015). Many in vivo end points cannot be predicted any better by using HTS
assay data than by using chemical descriptor information and QSAR models. Thomas et al. (2012) identified several factors that could account for the relatively poor ability of in vitro assays to predict in vitro
responses, including (1) the failure of current in vitro assays to capture biochemical and cellular processes
or properties in the in vivo tissues with adequate fidelity; (2) the possibility that in vitro assays do not
capture biological context-specific outcomes reliably; (3) inadequate coverage of pathways, protein targets, and cell types; (4) substantial species differences between the in vitro assays and in vivo end points
that are being predicted; and (5) the inadequate number of positive chemicals for each end point to capture the broad array of mechanisms that lead to in vivo toxicity.
98
Prepublication Copy
Integration and
BO
OX 6-3 The Us
se of HTS Ass
says for Identtifying Endoc rine Disruptin
ng Potential (C
Casey 2014)
In vivo phenotype:
p
Estrogenic bio
oactivity was assessed
a
witth the EPA an
nd Organisation for Econo
omic
Co-operration and De
evelopment (O
OECD) rodentt uterotrophicc assay in the
e ovariectomizzed rat, imma
ature
rat, or ovariectomize
o
d mouse. Duration of dosing (oral, sub
bcutaneous, o
or intraperiton
neal) was a m
minimum of 3 days. Suita
able in vivo da
ata on more th
han 100 chem
micals were id
dentified.
In vitro data: Results
s were obtain
ned from 18 in vitro assa
ays that meassure estrogen
n receptor (E
ER)
mediate
ed bioactivity. The in vitro assays
a
(chose
en from the E
EPA ToxCast or Tox21 pro
ogram) probe perturbation
ns of the ER pathway at multiple
m
points
s (ER binding
g, receptor dim
merization, DNA binding, R
RNA
transcrip
ption, protein production, and cell prolliferation) (Ro
otroff et al. 2
2014). Chemiicals used in the
evaluatio
on included several
s
reference chemicalls of known a ctivity againsst the ER (succh as estradio
ol).
Outputs: A computattional model was develop
ped to calcula
ate area-unde
er-the-curve ((AUC) scoress for
ER agonist (R1) and
d antagonist (R2) bioactivity. The scorres were com
mpared with tthe results off the
uterotrophic assay re
esults.
Conclus
sion: Data an
nalyses showed good corrrelation betw
ween the calcculated AUC scores (R1) and
results of
o the uterotro
ophic assay. The study de
emonstrated that ToxCastt in vitro assa
ays perform a
adequately for prioritizing
g chemicals for
f further eva
aluation of ER
R activity, and the HTS asssays are pre
edictive of th
he likelihood of
o a positive or
o negative fin
nding in more
e resource-inte
ensive assayys.
Cytotoxicity
C
Assays
A
as Geeneral Indicaators of Acutee Toxicity
As discussed
d
in Chapter
C
4, on
ne approach to
t predicting acute toxicityy relies on inn vitro cytotooxicity
assays thaat use human
n cells in cultture (Seibert et al. 1996; E
Ekwall et al. 1998; Halle 2003; NICE
EATM
2006; Xiaa et al. 2008; Halwachs et al. 2013; Prieeto et al. 20133a). That appproach uses reelatively simpple assays and presumes
p
thatt in vivo toxiccity does not result primarrily from imppairment of sppecific functioons of
differentiaated cells (Priieto et al. 201
13a). In many
y cases, validaation of the asssays rests onn an examination of
the regresssion equatio
on from the correlation
c
of experimenttal IC50 6 cytootoxicity valuues with publlished
LD50s,7 an
n approach th
hat can be used to estimaate unknown LD50 valuess from IC50 ccytotoxicity vvalues
measured in vitro (Seib
bert et al. 199
96). The com
mmittee examiined the NICE
EATM and E
ECVAM validdation
study thatt assessed thee predictive capacity
c
of th
he BALB/3T33 neutral-redd uptake cytottoxicity assayy (see
Box 6-4); that assay haas been evalu
uated for its ab
bility to identtify chemicals that would be labeled ass toxic
6
7
IC50 is the concentratiion of a substaance that causess 50% inhibitioon in vitro.

In som
me cases, LD50 values
v
are avaiilable from the Registry of Cyytotoxicity (Haalle 2003).
99
Appliccation of Mod
ous (LD50 < 2,000
2
mg/kg)) (NTP 2014)). The overalll results of thhe studies havve shown thatt there
or hazardo
is a relatively good co
orrelation of around
a
6070
0% between iin vitro cytottoxic concenttrations (IC50ss) and
rat oral LD
D50s (JRC 2013).
BOX 6-4
6 The Use of
o the BALB/3
3T3 Neutral-R
Red Uptake Cyytotoxicity Asssay to Predicct Acute Toxiccity
Assay: The
T 3T3 neuttral-red uptake
e (NRU) cytottoxicity assayy uses the BA
ALB/c3T3 mou
use fibroblastt cell
line and is based on the
t ability of viable cells to incorporate
i
an
nd bind the dyye neutral red
d (NR). The asssay
is usually performed as
a a 96-well cytotoxicity-ba
c
ased assay tha
at spectropho
otometrically m
measures (Sto
okes
et al. 2008) the conce
entration-depe
endent reduction in NRU byy cells after exxposure to a test material. The
basic co
oncept of basa
al cytotoxicity
y assays is that chemicals exert their to
oxic effects byy disrupting sttructures an
nd functions universal to all cells, such as
a cell membranes (Genna
ari et al. 2004
4). With the b
basal
cytotoxic
city assays, itt is possible to
t quantify the
e cytotoxicity of a chemica
al by its IC50 vvalue, that is,, the
concentrration of the te
ested substan
nce that decre
eases cell viab
bility by 50% iin the cell cultture.
Chemica
als: Chemicals tested in th
his assay inclluded pharma
aceuticals, pe
esticides, indu
ustrial chemiccals,
and food
d additives. The
T number off chemicals varied
v
betwee
en test phasess and ranged from about 7
70 to
300.
Predictiv
ve approach: The study assessed
a
the predictive ca
apacity of the
e assay in co
onjunction wiith a
dichotom
mous predictiion model tha
at yielded on
nly two categ
gorical predicttions: potential negative ((predicted LD50 > 2,000 mg/kg)
m
and po
otential positiv
ve (predicted LD50 < 2,000
0 mg/kg).
The figu
ure below (fro
om NICEATM 2006, P. 6-19) shows the
e regressionss that used the Registry off Cytotoxicity
y (RC; Halle 2003) rat acute oral LD50 data on mill imolar (LEFT
T) or weight ((RIGHT) unitss for
282 testt substances.
Overall assay
a
perform
mance: The 3T3
3 NRU method was show
wn to have a high sensitivity (9296%) and
a low false-negative rate (48%) (Prieto
(
et al. 2013a).
2
The a
assay also ha
ad a high falsse-positive ratte or
low spec
cificity, which limited the usefulness of positive test rresults and le
ed to a compa
arably low ratte of
identifica
ation of true negatives
n
as such
s
(4044%
%). However, negative tesst results of th
he 3T3 NRU w
were
largely accurate,
a
thatt is, substance
es identified as
a negatives had low toxiccity (low false--negative rate
e).
Assay liimitations: The 3T3 NRU assay
a
address
ses specificallly the toxicityy mechanism of basal cyto
otoxicity; fibrroblast cells cannot
c
be use
ed to evaluatte interactionss of chemicalls with neuron
nal or cardiacc receptors and ion chan
nnels and other tissue-spe
ecific molecula
ar targets (NIEHS 2009). Furthermore,, the
cell line lacks metabo
olic competen
nce associate
ed with Phase
e I and Phase
e II biotransforrmation and sso is
sensitive
e to cytotoxicity induced by
y the parent chemical
c
and not its metab
bolites.
100

Although cytotoxicity assays hold some promise for the prediction of acute toxicity, several important caveats and assay limitations need to be considered, including the following:
Little evidence of assay performance exists for highly toxic chemicals. DOD is faced with identifying agents that have extremely low LD50s or LC50s. For example, acute oral LD50 values reported in
mice for some agents of current concern to DOD are well below 1 mg/kg: botulinum toxin (0.001 g/kg),
ricin (3 g/kg), VX (15 g/kg), anatoxina(s) (50 g/kg), soman (64 g/kg), and sarin (100 g/kg).8 In
addition, those examples of highly toxic chemicals have different mechanisms of action, including inhibition of cholinesterase activity (for example, anatoxin-a(s), VX, soman, and sarin), ribosome inactivation
(ricin), and blockade of acetylcholine secretion (botulinum toxin). Applying the BALB/3T3 NRU cytotoxicity assay to a set of 67 chemicals and using the rat LD50 data from the Registry of Cytotoxicity
showed that predictions for highly toxic chemicals were generally poor0/6 for chemicals with an LD50
below 5 mg/kg and 1/11 chemicals with an LD50 of 550 mg/kg (NICEATM 2006). Substances with
LD50s of 3002,000 mg/kg were predicted better by this assayabout 81% accuracy (NICEATM 2006).
Prediction of acute toxicity with cytotoxicity assays remains highly variable. There are several
possible reasons why acute systemic toxicity of some chemicals is poorly predicted by basal cytotoxicity
assays. First, toxicity might be due to tissue- or organ-specific effects caused by the chemical of interest.
For example, substances with a mechanism of action on molecular targets (receptors, channels, and enzymes) are not modeled in most cell lines that are used for cytotoxicity assays, and this probably accounts
for poor prediction of the toxicity of anatoxin, botulinum toxin, soman, and sarin, which perturb neuronal
synapses. Second, even though a generally toxic mechanism is modeled by a cell line, the cell line could
be much less sensitive to this mechanism than is some sensitive cell type in the human body. For example, ricin targets protein synthesis, which is needed by all cells, but in the human body, only specific organs are highly sensitive. Third, chemicals have restricted accessibility to target cells or tissues. Fourth,
chemical toxicity might depend on bioactivation pathways that are absent in the test system. Fifth, a
chemical might be quickly eliminated or detoxified through metabolism; thus, in vitro results might overpredict toxicity seen in vivo.
Assay limitations can also contribute to reduced predictive power of in vitro cytotoxicity assays.
For example, actual concentrations available to cells or to intracellular targets in in vitro tests might be
much lower than the nominal concentrations.
The ACuteTox project identified several important technical considerations in the design of an in
vitro testing strategy (Kopp-Schneider et al. 2013). Some of the more important considerations were the
following:
Select chemicals to span a wide range of outcomes of interest.
Perform all assays with all chemicals.
Be aware that chemical solubility might limit the concentrations that can be used in the test system.
Design experiments carefully to guarantee reliable and meaningful estimates.
Avoid overfitting of the prediction model, which occurs when a model is fitted exactly to the training data; this hinders performance in future screening efforts or applications.
Use cross-validation and bootstrapping to estimate error rates.
A QSAR model based on in vitro cytotoxicity data and oral LD50 values from the Registry of Cytotoxicity has been developed for use in predicting acute toxicity in rodents (Freidig et al. 2007). The predictions from that model tend to overestimate toxicity; thus, substances that were predicted to have no or
low toxicity with that model could be eliminated from additional in vivo testing.
8
LD50 data available from Franz (1997).
Prepublication Copy
101
Mechanistically Based Assays
Another approach that has been used to predict acute chemical toxicity is based on the development
of assays that evaluate a mechanism of action known to be critical for a chemicals toxic response. One
mechanism-based example explored by the committee uses HTS assays and computational approaches to
predict chemical-induced inhibition of AChE activity. Inhibition of AChE results in an accumulation of
acetylcholine (ACh) at cholinergic synapses and associated clinical signs. A variety of agents, including
nerve agents (such as VX and soman) and other toxic organophosphorus (OP) chemicals and some
carbamates, inhibit AChE activity. Several molecular models have been developed that use reference
chemicals to describe inhibitor binding with human (and other) AChE (Barril et al. 2001; Kua et al. 2002;
Guo et al. 2004; Akula et al. 2006; Sopkova de Oliveira Santos et al. 2010; Gupta and Mohan 2011; Deb
et al. 2012). The models evaluate inhibitor interactions at the two principal binding sitescatalytic
anionic site (CAS) and a peripheral anionic site (PAS)in the AChE enzyme. The nerve gases and other
classical AChE inhibitors bind a phosphoryl group on a serine residue in the CAS (Deb et al. 2012).
Another approach for assessing chemicalAChE interactions is to evaluate how the chemical of interest
docks with the active site and other portions of the protein. Several structureactivity relationships
(SARs) based on pose9 predictions for the interactions have been developed (Huang et al. 2010; Samadi et
al. 2010). Results obtained with those computational approaches often depend on the type of ligand, the
protein conformation, and the presence of water (Berg et al. 2011). Box 6-5 illustrates several examples
of how an HTS approach can be used to screen chemicals as AChE inhibitors.
Similar mechanism-based screening efforts have been developed by the pharmaceutical industry and
others to develop in vitro models that can predict in vivo biology in support of drug-safety assessment or
drug discovery. For example, some investigators have examined the relationship between cytotoxicity
data, basal gene-expression measurements, and a chemicals structure to identfy putative molecular
targets and the role of gene expression in cytotoxicity (Covell et al. 2005). Another example involves
screening chemicals for their ability to inhibit mitochondrial complex I of the electron transport chain
(Glover et al. 2007). Berg et al. (2006) successfully grouped chemicals by mechanism of action (for
example, modulators of NFkB signaling or the phosphatidylinositol 3-kinase/Akt signal-transduction
pathway) by using data from human cell-based HTS assays. Such approaches are consistent with the
development of an adverse-outcome pathway (AOP), which is a linear conceptual model that links a
molecular initiating event, key events, and an adverse outcome (Figure 3-1) (Ankley et al. 2010; GarciaReyero 2015).
An AOP begins with a well-defined molecular initiating event and then describes the key events
along a biological pathway that ultimately lead to an adverse outcome associated with chemical exposure.
One has been developed for lethality associated with AChE -inhibiting organophosphate and carbamate
pesticides (Russom et al. 2014). That AOP considers the role of metabolism (such as metabolic activation
by cytochrome P450s to form oxon metabolites of some organophosphate pesticides) and links the main
initiating event (cholinesterase inhibition) with various physiological cholinergic responses. Whether that
or other AOPs in development (such as the ones for nonpolar narcosis and mitochondrial toxicity10) will
be relevant for the highly toxic chemicals of concern to DOD is unknown but deserves further study.
OTHER CONSIDERATIONS
The committee has identified several broad technical considerations that are important, such as the
use of reference chemicals and development of appropriate data-integration models. DOD is faced with
The term pose refers to the conformation and alignment of a molecule (Coats 2002).
OECD provides additional information about developing AOPs at http://www.oecd.org/chemicalsafety/testing/
listsofprojectsontheaopdevelopmentprogrammeworkplan.htm.
10
102
Prepublication Copy
Integration and
BOX 6-5 The Use of HTS

H
Assays to
o Evaluate In hibition of Accetylcholineste
erase
Berg et al. (2011) scrreened a chemical library consisting
c
of 17,500 substtances by usiing the colorim
metric Ellman assay adapted to a 96-well
9
format and a reco
ombinant hum
man AChE. T
The hydrolysis of
acetylthiocholine iodide was mon
nitored, and the
t
average slope of the
e positive con
ntrols was se
et to
100% activity. At an assay conce
entration of 50
0 M, 124 ch
hemicals redu
uced the enzzymatic activitty of
AChE by at least 70% in the sing
gle-replicate assays.
a
The cchemicals ha
ad a full dose
eresponse cu
urve
determin
ned to verify chemical inhibitory activity
y and identifyy potential false-positive rresults. A seccond
set of ch
hemicals thatt had no activ
vity in the HTS
S but structurral and physiccochemical fe
eatures simila
ar to
those off the positive hits was use
ed to identify false negativves. The ACh
hE inhibitors d
discovered in
n the
screenin
ng campaign are chemically diverse, wiith molecular weights rang
ging from 234
4 to 596 Da, llogP
(o/w) fro
om 1.16 to 8.14,
8
and 01
12 rotatable bonds.
b
Five p
principal com
mponents (PC
Cs) proved sig
gnificant: the
ey were relate
ed mainly to size,
s
hydroph
hobicity, flexib
bility, charge (positive, neu
utral or negative),
and elec
ctronic properrties associatted with halog
gens and aro matic elemen
nts (Berg et a
al. 2011). The
e figure belo
ow illustrates the
t chemical space as esttablished by P
PC analysis o
of the physico
ochemical pro
operties of th
he 124 hits (gray dots) thatt were identified in the HTS
S performed b
by Berg et al. (2011).
several ad
dditional challenges that will
w need to bee considered aas it developss a tiered testting program. They
include teechnical challlenges associaated with the assessment oof chemicals for direct debbilitating porttal-ofentry effeects (such as skin blisterin
ng and pulmo
onary edema)) because theere are few inn vitro system
ms for
assessing the acute tox
xicity of inhalled chemicalss or agents thaat produce skkin blistering. Likewise, it is unknown wh
hether toxicitty screens dev
veloped to prredict oral toxxicity could aalso predict ddermal LD50s or inhalation LC
L 50s. Anotheer important challenge
c
thatt DOD faces is related to tthe broad arraay of chemicaals (or
chemical classes) that could requiree assessment. That issue m
might require DOD to develop distinct tiered
nt chemical cllasses, such as
a metals. DO
OD will also nneed to support the developpment
approachees for differen
of HTS assays for mecchanisms of actions
a
that are
a known orr suspected too be involvedd in acute cheemical
toxicity. In
I the future, AOPs and QS
SARs could be
b developed to strengthenn DODs abillity to predict acute
chemical toxicity (Tolllefsen et al. 2014).
2
Finally
y, DOD will nneed to devellop methods ffor integratingg data
ysicochemicaal and biological domains.
across phy
DOD
D has a historry of using alternative
a
testt methods forr the assessm
ment of chemical-warfare aagents.
For examp
ple, the US Armys
A
Medical Chemical Defense
D
Reseearch Program
m used a varieety of assays tto elucidate meechanisms of action and identify
i
coun
ntermeasures for many off the classicall chemical-w
warfare
agents, su
uch as sulfur mustard,
m
lewissite, nerve ageents, and hydrrogen cyanidee (see Siddell eet al. 1997; Soomani
and Romaano 2001). Co
onsiderable DO
OD effort weent into being able to conduuct the assays safely when working with agents,
a
such as
a soman and VX, which have
h
human deermal LD50 esstimates of 0..14 and 5 g //kg of
body weig
ght, respectiveely. If the oveert toxicity of a new or novvel chemical is totally unknnown and withhin an
103
order of magnitude or so of a classical chemical-warfare agents potency, serious consideration of the engineering controls necessary to conduct the assays will be required. Because few research facilities are
equipped to work with known highly toxic chemical-warfare agents as reference chemicals, DOD might be
required initially to use surrogate chemicals that have a shared chemical mechanism or clinical effect of
more militarily relevant agents.
The committee expects that in the next 310 years any tiered testing approach will not be able to replace completely the need for follow-up targeted in vivo studies to confirm the toxicity of a chemical of
interest. Indeed, the state of the science suggests that development of a predictive acute-toxicity program
will require extensive DOD investment in development of fit-for-purpose assays, (Q)SAR and other computational modeling approaches, and in vitroin vivo extrapolation and data-integration methods for the
prediction of acute toxicity. To begin the investment, the committee recommends that DOD initiate pilot
studies that evaluate chemical classes of highest concern with well-characterized reference chemicals.
The pilot studies will allow it to develop the assays and tools that are needed to predict acute chemical
toxicity efficiently and accurately and to evaluate the rate of false negatives and false positives. The pilot
studies could also examine how generalizable the results of various assays and tools are from one chemical class11 to another. That research would allow DOD to begin to address the size of the chemical space
needed to make predictions about unknown chemicals. DOD could benefit from leveraging its efforts
with existing federal activities, such as the ToxCast program. Collaboration would allow DOD to complete pilot studies more rapidly.
Finding: On the basis of its review of the current state of the science, the committee concludes
that development of a screening program to predict acute toxicity that incorporates such information as
existing toxicity data, physicochemical properties, and results from in silico modeling and in vitro testing
is consistent with and supported by other testing frameworks that use modern toxicology data.
Finding: The current state of the science for prediction of acute toxicity with computational approaches is seeing steady growth. An investment by DOD in computational approaches could yield benefits in characterizing the toxicity of chemicals on which toxicity data are sparse or lacking.
Finding: Prediction of acute toxicity with HTS assays is in its infancy, and DOD will need to
evaluate what assays or approaches might be applicable for evaluating acute toxicity for its system and
consider the lessons learned from current large-scale HTS programs. Regardless, an investment by DOD
in HTS approaches could yield benefits in characterizing the toxicity of chemicals on which few or no
toxicity data are available. HTS approaches might prove useful in excluding chemicals that have low toxic potential (for example, rodent LD50 greater than 1,000 mg/kg) from further testing and might also help
to identify more toxic chemicals of concern for further testing.
Finding: There are sufficient data to suggest that DOD could use simple cytotoxicity assays to
identify chemicals that have low acute-toxicity potential; this would allow it to focus its attention on more
toxic chemicals. Additional effort is needed to determine whether the assays are relevant for the identification of highly toxic chemicals that could be used against deployed troops. Furthermore, because simple
cytotoxicity assays fail to predict toxicity of highly toxic chemicals that act on specific molecular targets,
such as neuronal synapses, they would need to be supplemented with assays designed specifically to detect such effects.
Finding: On the basis of scientific advances, the committtee concludes that the development of
appropriate cellular models and targeted mechanistically based assays could provide DOD with a useful
11
In this context, chemical class is defined broadly to include structurally related chemicals, chemicals that have
different mechanisms of action, and chemicals that have different toxic end points, such as hepatotoxicity and neurotoxicity.
104
Prepublication Copy

resource for understanding and predicting potential toxicity of chemicals. In particular, having more explicit knowledge available on the numerous mechanisms of action that lead to acute systemic toxicity
would be valuable in the design and validation of integrated prediction methods.
Recommendation: DOD should initiate pilot studies that evaluate chemical classes of highest
concern with well-characterized reference chemicals. The pilot studies will allow it to develop the assays
and tools that are needed to predict acute chemical toxicity efficiently and accurately and to evaluate the
rate of false negatives and false positives.
REFERENCES
Akula, N., L. Lecanu, J. Greeson, and V. Papadopoulos. 2006. 3D QSAR studies of AChE inhibitors based on molecular docking scores and CoMFA. Bioorg. Med. Chem. Lett. 16(24):6277-6280.
Ankley, G.T., R.S. Bennett, R.J. Erickson, D.J. Hoff, M.W. Hornung, R.D. Johnson, D.R. Mount, J.W. Nichols,
C.L. Russom, P.K. Schmieder, J.A. Serrrano, J.E. Tietge, and D.L. Villeneuve. 2010. Adverse outcome pathways: A conceptual framework to support ecotoxicology research and risk assessment. Environ. Toxicol.
Chem. 29(3):730-741.
Barril, X., M. Orozco, and F.J. Luque. 2001. Towards improved acetylcholinesterase inhibitors:A structural and
computational approach. Mini Rev. Med. Chem. 1(3):255-266.
Becker, R.A., K.P. Friedman, T.W. Simon, M.S. Marty, G. Patlewicz, and J.C. Rowlands. 2015. An exposure: Activity profiling method for interpreting high-throughput screening data for estrogenic activity-Proof of concept. Regul. Toxicol. Pharmacol. 71(3):398-408.
Berg, L., C.D. Andersson, E. Artursson, A. Hrnberg, A.K. Tunemalm, A. Linusson, and F. Ekstrm. 2011. Targeting acetylcholinesterase: Identification of chemical leads by high throughput screening, structure determination and molecular modeling. PLoS One 6(11):e26039.
Berg, E.L., E.J. Kunkel, E. Hytopoulos, and I. Plavec. 2006. Characterization of compound mechanisms and secondary activities by BioMAP analysis. J. Pharmacol. Toxicol. Meth. 53(1):67-74.
Can, A. 2014. Quantitative structure-toxicity relationship (QSTR) studies on the organophosphate insecticides. Toxicol. Lett. 230(3):434-443.
Casey, W. 2014. Assessing the Biological Relevance of In Vitro Data: A Case Study Using Estrogen Pathway Signaling. Presentation at NTP Board of Scientific Counselors (BSC) Meeting, December 9-10, 2014 [online]. Available: http://ntp.niehs.nih.gov/ntp/about_ntp/bsc/2014/dec/presentations/04casey_bscdec2014_508.pdf [accessed
April 6, 2015].
Chen, M., C.W. Tung, Q. Shi, L. Guo, L. Shi, H. Fang, J. Borlak, and W. Tong. 2014a. A testing strategy to predict
risk for drug-induced liver injury in humans using high-content screen assays and the rule-of-two model.
Arch. Toxicol. 88(7):1439-1449.
Chen, M., H. Bisgin, L. Tong, H. Hong, H. Fang, J. Borlak, and W. Tong. 2014b. Toward predictive models for
drug-induced liver injury in humans: Are we there yet? Biomark. Med. 8(2):201-213.
Clemedson, C. 2008. The European ACuteTox project: A modern integrative in vitro approach to better prediction
of acute toxicity. Clin. Pharmacol. Ther. 84(2):200-202.
Clothier, R., P. Dierickx, T. Lakhanisky, M. Fabre, M. Betanzos, R. Curren, M. Sjstrm, H. Raabe, N. Bourne, V.
Hernandez, J. Mainez, M. Owen, S. Watts, and R. Anthonissen. 2008. A database of IC50 values and principal
component analysis of results from six basal cytotoxicity assays, for use in the modelling of the in vivo and in
vitro data of the EU ACuteTox project. ATLA 36(5):503-519.
Coats EA. Part III. 3D QSAR applications. In: 3D QSAR in Drug Design: Volume 3: Recent Advances. Edited by
Kubinyi H, Folkers G, Martin YC, 2002, Kluwer Academic Publishers, New York, NY, pp 197-231.
Covell, D.G., A. Wallqvist, R. Huang, N. Thanki, A.A. Rabow, and X.J. Lu. 2005. Linking tumor cell cytotoxicity
to mechanism of drug action: An integrated analysis of gene expression, small-molecule screening and structural databases. Proteins 59(3):403-433.
Cox, L.A., D. Popken, M.S. Marty, J.C. Rowlands, G. Patlewicz, K.O. Goyak, and R.A. Becker. 2014. Developing
scientific confidence in HTS-derived prediction models: Lessons learned from an endocrine case study. Regul.
Toxicol. Pharmacol. 69(3):443-450.
Deb, P.K., A. Sharma, P. Piplani, and R.R. Akkinepally. 2012. Molecular docking and receptor-specific 3D-QSAR
studies of acetylcholinesterase inhibitors. Mol. Divers. 16(4):803-823.
Prepublication Copy
105
Dix, D.J., K.A. Houck, M.T. Martin, A.M. Richard, R.W. Setzer, and R.J. Kavlock. 2007. The ToxCast program for
prioritizing toxicity testing of environmental chemicals. Toxicol. Sci. 95(1):5-12.
Ekwall, B., F.A. Barile, A. Castao, C. Clemedson, R.H. Clothier, P. Dierickx, B. Ekwall, M. Ferro, G. Fiskesj, L.
Garza-Ocaas, M.J. Gmez-Lechn, M. Glden, T. Hall, B. Isomaa, A. Kahru, G. Kerszman, U. Kristen, M.
Kunimoto, S. Krenlampi, L. Lewan, A. Loukianov, T. Ohno, G. Persoone, L. Romert, T.W. Sawyer, R.
Shrivastava, H. Segner, A. Stammati, N. Tanaka, M. Valentino, E. Walum, and F. Zucco. 1998. MEIC evaluation of acute systemic toxicity. Part VI. The prediction of human toxicity by rodent LD50 values and results
from 61 in vitro methods. ATLA 26(Suppl. 2):617-658.
Franz, D.R. 1997. Defense against Toxin Weapons. US Army Medical Research and Materiel Command [online].
Available: http://www.usamriid.army.mil/education/defensetox/toxdefbook.pdf [accessed December 24, 2014].
Freidig, A.P., S. Dekkers, M. Verwei, E. Zvinavashe, J.G. Bessems, and J.J. van de Sandt. 2007. Development of a
QSAR for worst case estimates of acute toxicity of chemically reactive compounds. Toxicol. Lett. 170(3):214222.
Garcia-Reyero, N. 2015. Are adverse outcome pathways here to stay? Environ. Sci. Technol. 49(1):3-9.
Gennari, A., C. van den Berghe, S. Casati, J. Castell, C. Clemedson, S. Coecke, A. Colombo, R. Curren, G. Dal Negro, A. Goldberg, C. Gosmore, T. Hartung, I. Langezaal, I. Lessigiarska, W. Maas, I. Mangelsdorf, R. Parchment, P. Prieto, J. Riego Sintes, M. Ryan, G. Schmuck, K. Stitzel, W. Stokes, J.A. Vericat, and L. Gribaldo.
2004. Strategies to replace in vivo acute systemic toxicity testing. The report and recommendations of
ECVAM Workshop 50. ATLA 32(4):437-459.
Glover, C.J., A.A. Rabow, Y.G. Isgor, R.H. Shoemaker, and D.G. Covell. 2007. Data mining of NCI's anticancer
screening database reveals mitochondrial complex I inhibitors cytotoxic to leukemia cell lines. Biochem.
Pharmacol. 73(3):331-340.
Guo, J., M.M. Hurley, J.B. Wright, and G.H. Lushington. 2004. A docking score function for estimating ligandprotein interactions: Application to acetylcholinesterase inhibition. J. Med. Chem. 47(22):5492-5500.
Gupta, S., and C.G. Mohan. 2011. 3D-pharmacophore model based virtual screening to identify dual-binding site
and selective acetylcholinesterase inhibitors. Med. Chem. Res. 20(9):1422-1430.
Halle, W. 2003. The registry of cytotoxicity: Toxicity testing in cell cultures to predict acute toxicity (LD50) and to
reduce testing in animals. ATLA 31(2):89-198.
Halwachs, S., C. Lakoma, and W. Honscha. 2013. Validation of the hepatocyte-like HPCT-1E3 cell line as an in
vitro model for the prediction of acute in vivo toxicity. ATLA 41(5):369-383.
Huang, L., A.D. Shi, F. He, and X.S. Li. 2010. Synthesis, biological evaluation, and molecular modeling of berberine derivatives as potent acetylcholinesterase inhibitors. Bioorg. Med. Chem. 18(3):1244-1251.
Huang, R., S. Sakamuru, M.T. Martin, D.M. Reif, R.S. Judson, K.A. Houck, W. Casey, J.H. Hsieh, K.R. Shockley,
P. Ceger, J. Fostel, K.L. Witt, W. Tong, D.M. Rotroff, T. Zhao, P. Shinn, A. Simeonov, D.J. Dix, C.P. Austin,
R.J. Kavlock, R.R. Tice, and M. Xia. 2014. Profiling of the Tox21 10K compound library for agonists and antagonists of the estrogen receptor alpha signaling pathway. Sci. Rep. 4:5664.
JRC (Joint Research Centre). 2013. EURL ECVAM Recommendation on the 3T3 Neutral Red Uptake Cytotoxicity
Assay for Acute Oral Toxicity Testing. Report EUR 25946. Luxembourg: Office of European Union [online].
Available: https://eurl-ecvam.jrc.ec.europa.eu/eurl-ecvam-recommendations/files-3t3/ReqNo_JRC79556_lbna
25946enn.pdf [accessed December 24, 2014].
Judson, R.S., K.A. Houck, R.J. Kavlock, T.B. Knudsen, M.T. Martin, H.M. Mortensen, D.M. Reif, D.M.Rotroff, I.
Judson, R., K. Houck, M. Martin, T. Knudsen, R.S. Thomas, N. Sipes, I. Shah, J. Wambaugh, and K. Crofton. 2014.
In vitro and modelling approaches to risk assessment from the US Environmental Protection Agency ToxCast
programme. Basic Clin. Pharmacol. Toxicol. 115(1):69-76.
Kavlock, R., K. Chandler, K. Houck, S. Hunter, R. Judson, N. Kleinstreuer, T. Knudsen, M. Martin, S. Padilla, D.
Reif, A. Richard, D. Rotroff, N. Sipes, and D. Dix. 2012. Update on EPA's ToxCast program: Providing high
Kinsner-Ovaskainen, A., R. Rzepka, R. Rudowski, S. Coecke, T. Cole, and P. Prieto. 2009. Acutoxbase, an innovative database for in vitro acute toxicity studies. Toxicol In Vitro 23(3):476-485.
Kinsner-Ovaskainen, A., P. Prieto, S. Stanzel, and Kopp-Schneider. 2013. Selection of test methods to be included
in a testing strategy to predict acute oral toxicity: An approach based on statistical analysis of data collected in
phase 1 of the ACuteTox project. Toxical. In Vitro 27(4):1377-1394.
106
Prepublication Copy

Kleinstreuer, N.C., ,J. Yang,, E.L Berg, T.B. Knudsen, A.M Richard, M.T Martin, D.M Reif, R.S Judson, M. Polokoff, D.J Dix, R.J Kavlock, and K.A Houck. 2014. Phenotypic screening of the ToxCast chemical library to
classify toxic and therapeutic mechanisms. Nat. Biotechnol. 32(6):583-591.
Kopp-Schneider, A., P. Prieto, A. Kinsner-Ovaskainen, and S. Stanzel. 2013. Design of a testing strategy using nonanimal based test methods: Lessons learnt from the ACuteTox project. Toxicol. In Vitro 27(4):1395-1401.
Kua, J., Y. Zhang, and J.A. Mccammon. 2002. Studying enzyme binding specificity in acetylcholinesterase using a
combined molecular dynamics and multiple docking approach. J. Am. Chem. Soc. 124(28):8260-8267.
Lipinski, C., and A. Hopkins. 2004. Navigating chemical space for biology and medicine. Nature 432(7019):855861.
Low, Y., T. Uehara, Y. Minowa, H. Yamada, Y. Ohno, T. Urushidani, A. Sedykh, E. Muratov, D. Fourches, H. Zhu,
I. Rusyn, and A. Tropsha. 2011. Predicting drug-induced hepatotoxicity using QSAR and toxicogenomics approaches. Chem. Res. Toxicol. 24(8):1251-1262.
NICEATM (National Toxicology Program Interagency Center for the Evaluation of Alternative Toxicological
Methods), 2006. In Vitro Cytotoxicity Test Methods for Estimating Acute Oral Systemic Toxicity. Background Review Document (BRD). NIH Publication No. 07-4518. National Institute of Health, National Institute of Environmental Health Sciences, Research Triangle Park, NC. November 2006 [online]. Available:
http://ntp.niehs.nih.gov/pubhealth/evalatm/test-method-evaluations/acute-systemic-tox/in-vitrovalidation/brd/index.html [accessed March 26, 2015].
NIEHS (National Institute of Environmental Health Sciences). 2009. Report on the ICCVAM-ICEATM/ECVAM/
JaCVAM Scientific Workshop on Acute Chemical Safety Testing: Advancing In Vitro Approaches and Humane Endpoints for Systemic Toxicity Evaluations. Interagency Coordinating Committee on the Validation of
Alternative Methods (ICCVAM), National Toxicology Program Interagency Center for the Evaluation of Alternative Toxicological Methods (NICEATM), National Institute of Environmental Health Sciences, National
Toxicology Program, Research Triangle Park, NC [online]. Available: http://ntp.niehs.nih.gov/iccvam/docs/
acutetox_docs/workshop-feb09/acutewksprpt.pdf [accessed April 6, 2015].
NTP (National Toxicology Program). 2014. Validation Study of In Vitro Cytotoxicity Test Methods [online].
Available: http://ntp.niehs.nih.gov/pubhealth/evalatm/test-method-evaluations/acute-systemic-tox/in-vitro-v
alidation/ [accessed April 6, 2015].
Olson, H., G. Betton, D. Robinson, K. Thomas, A. Monro, G. Kolaja, P. Lilly, J. Sanders, G. Sipes, W. Bracken, M.
Dorato, K. Van Deun, P. Smith, B. Berger, and A. Heller. 2000. Concordance of the toxicity of pharmaceuticals in humans and in animals. Regul. Toxicol. Pharmacol. 32(1):56-67.
Patlewicz, G., T.W. Simon, J.C. Rowlands, R.A. Budinsky, and R.A. Becker. 2015. Proposing a scientific confidence framework to help support the application of adverse outcome pathways for regulatory purposes. Regul.
Toxicol. Pharmacol. 71(3):463-477.
Prieto, P., T. Cole, R. Curren, R.M. Gibson, M. Liebsch, H. Raabe, A.M. Tuomainen, M. Whelan, and A. KinsnerOvaskainen. 2013a. Assessment of the predictive capacity of the 3T3 Neutral Red Uptake cytotoxicity test
method to identify substances not classified for acute oral toxicity (LD50>2000 mg/kg): Results of an
ECVAM validation study. Regul. Toxicol. Pharmacol. 65(3):344-365.
Prieto, P., A. Kinsner-Ovaskainen, S. Stanzel, B. Albella, P. Artursson, N. Campillo, R. Cecchelli, L. Cerrato, L.
Daz, E. Di Consiglio, A. Guerra, L. Gombau, G. Herrera, P. Honegger, C. Landry, J.E. OConnor, J.A. Pez,
G. Quintas, R. Svensson, L. Turco, M.G. Zurich, M.J. Zurbano, and A. Kopp-Schneider. 2013b. The value of
selected in vitro and in silico methods to predict acute oral toxicity in a regulatory context: Results from the
European Project ACuteTox. Toxical. In Vitro 27(4):1357-1376.
Rotroff, D.M., D.J. Dix, K.A. Houck, T.B. Knudsen, M.T. Martin, K.W. McLaurin, D.M. Reif, M. Crofton, A.V.
Singh, M. Xia, R. Huang, and R.S. Judson. 2013. Using in vitro high throughput screening assays to identify
potential endocrine disrupting chemicals. Environ. Health Perspect. 121(1):7-14.
Rotroff, D.M., D.J. Dix, K.A. Houck, T.B. Knudsen, M.T. Martin, K.W. McLaurin, D.M. Reif, K.M. Crofton, A.V.
Singh, M. Xia, R. Huang, and R.S. Judson. 2014. Using in vitro high throughput screening assays to identify
potential endocrine-disrupting chemicals. Environ. Sci. Technol. 48(15):8706-8716.
Russom, C.L., C.A. LaLone, D.L. Villeneuve, and G.T. Ankley. 2014. Development of an adverse outcome pathway
for acetylcholinesterase inhibition leading to acute mortality. Environ. Toxicol. Chem. 33(10):2157-2169.
Rusyn, I., A. Sedykh, Y. Low, K.Z. Guyton, and A. Tropsha. 2012. Predictive modeling of chemical hazard by integrating numerical descriptors of chemical structures and short-term toxicity assay data. Toxicol. Sci. 127(1):1-9.
Samadi, A., J. Marco-Contelles, E. Soriano, M. lvarez-Prez, M. Chioua, A. Romero, L. Gonzlez-Lafuente, L.
Ganda, J.M. Roda, M.G. Lpez, M. Villarroya, A.G. Garca, and C. de los Ros. 2010. Multipotent drugs with
cholinergic and neuroprotective properties for the treatment of Alzheimer and neuronal vascular diseases. I.
Prepublication Copy
107
Synthesis, biological assessment, and molecular modeling of simple and readily available 2-aminopyridine-,
and 2-chloropyridine-3,5-dicarbonitriles. Bioorg. Med. Chem. 18(16):5861-5872.
Seibert, H., M. Balls, J.H. Fentem, V. Bianchi, R.H. Clothier, P.J. Dierickx, M. Garle, M.J. Gomez-Lechon, L.
Gribaldo, M. Gulden, M. Liebsch, E. Rasmussen, R. Roguet, R. Shrivastana, and E. Walum, 1996. Acute toxicity testing in vitro and the classification and labelling of chemicals. The report and recommendations of
ECVAM Workshop 16. ATLA 24:499-510.
Seidle, T., S. Robinson, T. Holmes, S. Creton, P. Prieto, J. Scheel, and M. Chlebus. 2010. Cross-sector review of
drivers and available 3Rs approaches for acute systemic toxicity testing. Toxicol. Sci. 116(2):382-396.
Siddell, F.R., E.T. Takafuji, and D.R. Franz, eds. 1997. Medical Aspects of Chemical and Biological Warfare. Textbook of Medicine, Part 1. Warfare, Weaponry and the Casualty, Vol. 3. Office of Surgeon General, Borden Institute, Walter Reed Army Medical center, Washington, DC.
Sipes, N.S., M.T. Martin, P. Kothiya, D.M. Reif, R.S. Judson, A.M. Richard, K.A. Houck, D.J. Dix, R.J. Kavlock,
and T.B. Knudsen. 2013. Profiling 976 ToxCast chemicals across 331 enzymatic and receptor signaling assays. Chem. Res. Toxicol. 26(6):878-895.
Sjstrm, M., A. Kolman, C. Clemedson, and R. Clothier. 2008. Estimation of human blood LC50 values for use in
modeling of in vitro-in vivo data of the ACuteTox project. Toxicol. In Vitro 22(5):1405-1411.
Somani, S.M., and J.A. Romano, eds. 2001. Chemical Warfare Agents: Toxicity at Low Levels. Boca Raton, FL:
CRC Press.
Sopkova de Oliveira Santos, J., A. Lesnard, J.H. Agondanou, N. Dupont, A.M. Godard, S. Stiebing, C. Rochais, F.
Fabis, P. Dallemagne, R. Bureau, and S. Rault. 2010. Virtual screening discovery of new acetylcholinesterase
inhibitors issued from CERMN chemical library. J. Chem. Inf. Model. 50(3):422-428..
Stephens, C., R.J. Andrade, and M.I. Lucena. 2014. Mechanisms of drug-induced liver injury. Curr. Opin. Allergy
Clin. Immunol. 14(4):286-292.
Stokes, W.S., S. Casati, J. Strickland, and M. Paris. 2008. Neutral red uptake cytotoxicity tests for estimating starting doses for acute oral toxicity tests. Curr. Protoc. Toxicol. (Suppl. 36):Unit 20.4.
Stokes, W.S., J. Strickland, and W. Casey. 2012. Validation of the 21st century toxicology toolbox: Challenges, opportunities, and the way forward. Pp. 323-328 in Proceedings of the 8th World Congress on Alternatives and
Animal Use in the Life Sciences, Montreal 2011. ALTEX Proceedings Vol. 1 [online]. Available:
http://www.altex.ch/resources/WC8proceedings_full_issue.pdf [accessed March 27, 2015].
Thomas, R.S., M.B. Black, L. Li, E. Healy, T.M. Chu, W. Bao, M.E. Andersen, and R.D. Wolfinger. 2012. A comprehensive statistical analysis of predicting in vivo hazard using high-throughput in vitro screening. Toxicol.
Sci. 128(2):398-417.
Thomas, R.S., M.A. Philbert, S.S. Auerbach, B.A. Wetmore, M.J. Devito, I. Cote, J.C. Rowlands, M.P. Whelan,
S.M. Hays, M.E. Andersen, M.E. Meek, L.W. Reiter, J.C. Lambert, H.J. Clewell, III, M.L. Stephens, Q.J.
Zhao, S.C. Wesselkamper, L. Flowers, E.W. Carney, T.P. Pastoor, D.D. Petersen, C.L. Yauk, and A. Nong.
2013. Incorporating new technologies into toxicity testing and risk assessment: Moving from 21st century vision to a data-driven framework. Toxicol. Sci. 136(1):4-18.
Tice, R.R., C.P. Austin, R.J. Kavlock, and J.R. Bucher. 2013. Improving the human hazard characterization of
chemicals: a Tox21 update. Environ. Health Perspect. 121(7):756-765.
Tollefsen, K.E., S. Scholz, M.T. Cronin, S.W. Edwards, J. de Knecht, K. Crofton, N. Garcia-Reyero, T. Hartung, A.
Worth, and G. Patlewicz. 2014. Applying Adverse Outcome Pathways (AOPs) to support Integrated Approaches to Testing and Assessment (IATA). Regul. Toxicol. Pharmacol. 70(3):629-640.
van der Burg, B., E.B. Wedebye, D.R. Dietrich, J. Jaworska, I. Mangelsdorf, E. Paune, M. Schwarz, A.H. Piersma,
and E.D. Kroese. In press. The ChemScreen project to design a pragmatic alternative approach to predict reproductive toxicity of chemicals. Reprod. Toxicol.
Wind. M., and W.S. Stokes. 2010. Developing performance standards to expedite validation of innovative and improved test methods. ALTEX 27(Special Issue):97-102.
Xia, M., R. Huang, K.L. Witt, N. Southall, J. Fostel, M.H. Cho, A. Jadhav, C.S. Smith, J. Inglese, C.J. Portier, R.R.
Tice, and C.P. Austin. 2008. Compound cytotoxicity profiling using quantitative high-throughput screening.
Environ. Health Perspect. 116(3):284-291.
Xu, J.J., D. Diaz, and P.J. OBrien. 2004. Applications of cytotoxicity assays and pre-lethal mechanistic assays for
assessment of human hepatotoxicity potential. Chem. Biol. Interact. 150(1):115-128.
Zhu, X, A. Sedykh, and S. Liu. 2014. Hybrid in silico models for drug-induced liver injury using chemical descriptors and in vitro cell-imaging information J. Appl. Toxicol. 34(3):281-288.
108
Prepublication Copy
Appendix A
Biographical Information on the Committee
on Predictive-Toxicology Approaches for
Military Assessments of Acute Exposures
David C. Dorman (Chair) is a professor of toxicology in the Department of Molecular Biosciences of

North Carolina State University. The primary objective of his research is to provide a refined understanding of chemically induced neurotoxicity in laboratory animals that will lead to improved assessment of
potential neurotoxicity in humans. Dr. Dorman's research interests include neurotoxicology, nasal toxicology, pharmacokinetics, and cognition and olfaction in animals. He has served on multiple National
Research Council committees and recently chaired the Committee on Design and Evaluation of Safer
Chemical Substitutions: A Framework to Inform Government and Industry Decisions. He has served on
other advisory boards for the US Navy, the National Aeronautics and Space Administration, and the US
Department of Agriculture and is currently a member of the National Toxicology Programs Board of
Scientific Counselors. He is an elected fellow of the Academy of Toxicological Sciences and a fellow of
the American Association for the Advancement of Science. He received his DVM from Colorado State
University. He completed a combined PhD and residency program in toxicology at the University of Illinois at Urbana-Champaign and is a diplomate of the American Board of Veterinary Toxicology and the
American Board of Toxicology.
Weihsueh A. Chiu is a professor in the Department of Veterinary Integrative Biosciences in the College
of Veterinary Medicine and Biomedical Sciences at Texas A&M University. Before joining Texas A&M
University, Dr. Chiu worked at the US Environmental Protection Agency (EPA) for over 14 years, most
recently as chief of the Toxicity Pathways Branch in the Integrated Risk Information System (IRIS) Division of the National Center for Environmental Assessment. His research has focused on human health risk
assessment, particularly with respect to toxicokinetics, mechanisms of toxicity, physiologically based
pharmacokinetic modeling, doseresponse assessment, and characterizing uncertainty and variability. He
led the development of EPAs 2011 IRIS assessment of trichloroethylene, which pioneered the use of
probabilistic methods for characterizing uncertainty and variability in toxicokinetics and doseresponse
relationships. Dr. Chiu received his PhD in physics from Princeton University.
Haiyan Huang is an associate professor in the Department of Statistics at the University of California,
Berkeley. She is interested in the interface between statistics and data-rich scientific disciplines, such as
biology. Her research is focused on applied and theoretical statistics, high dimensional and integrative
genomic data analysis, hierarchical multilabel classification, and translational bioinformatics. Dr. Huang
earned a PhD in applied mathematics from the University of Southern California.
Andy Nong is the lead computational toxicologist of Health Canada. His research focuses on computer
models of biological systems that can be applied to understand and predict the fate of a chemical dose in
the body and its possible health effects. Dr. Nong also explores different types of computing approaches
(pharmacokinetic models, benchmark dosing, chemical structureactivity regression, and systems-biology
Prepublication Copy
109
models) that can help to evaluate chemical safety and eventually support the screening of a larger set of
chemicals that lead to similar health outcomes. Before coming to Health Canada, he worked as a research
investigator with the Hamner Institutes for Health Sciences. Dr. Nong received his PhD in public health
and toxicology from the University of Montreal.
Grace Patlewicz is a research chemist at the National Center for Computational Toxicology of the US
Environmental Protection Agency (EPA). Before joining EPA, she was employed as a computational toxicologist by DuPonts Haskell Global Centers for Health and Environmental Sciences where she acted as
a focal point and technical lead for all (quantitative) structureactivity relationships ([Q]SAR) and readacross queries for product stewardship and regulatory purposes. A chemist and toxicologist by training,
she started her career as a safety-evaluation scientist at Unilever before focusing her interests in computational toxicology and moving into a role that involved providing modeling and chemistry expertise for a
variety of projects. While working for the (Q)SAR group at the European Commissions Joint Research
Centre, she was involved in many activities related to the development of technical guidance for Registration, Evaluation, Authorisation, and Restriction of Chemicals (REACH), including investigation of the
feasibility of using computational approaches in the development of chemical categories, development
and evaluation of (Q)SAR models for human health, and coordination of the technical development of
software tools, such as Toxtree and Toxmatch. Dr. Patlewicz received her PhD in organic chemistry from
the University of Santiago de Compostela in Spain.
David M. Reif is an associate professor of biological sciences at North Carolina State University and resident member of the Bioinformatics Research Center. His research focuses on the complex interactions
between human health and the environment through the integrated analysis of high-dimensional data from
diverse sources, including epidemiological studies of human health, high-throughput screening of environmental chemicals, and model organism data. Dr. Reif was previously a principal investigator with the
US Environmental Protection Agencys National Center for Computational Toxicology, where he led
several statistical and bioinformatics efforts and collaborated on a variety of projects with federal, academic, and industry partners. Dr. Reif received his PhD in human genetics from Vanderbilt University.
John Wade is vice president and general manager of the Life Sciences Research Business in the National
Security Division of Battelle. He is responsible for all Battelles animal-use laboratories and activities, in
particular chemicalbiological defense research, development, test, and evaluation and general toxicology
test and evaluation services for both government and commercial customers. Dr. Wade has been a member-at-large of the NATO Human Factors and Medicine Panel as the US delegations chemical- and biological-defense expert since 2000. He also serves as the chairman emeritus of the National Defense Industrial Associations Chemical and Biological Defense Division. Dr. Wade received his DVM from
Michigan State University and his PhD in toxicology from the University of Kansas School of Medicine,
and he was a diplomate of the American Board of Toxicology from 1991 to 2006.
Katrina Waters is the deputy division director for biological sciences at the Pacific Northwest National
Laboratory. Her research interests are reconstruction of cell-response networks from integrated gene- and
protein-expression data to enable predictive mechanistic modeling of disease and toxicity pathways. Dr.
Waters serves on the US Food and Drug Administration National Center for Toxicological Researchs
Science Advisory Board and the US Environmental Protection Agencys Scientific Advisory Panel on
Methods for Prioritizing Endocrine Disrupting Chemicals. She is a member of the Society of Toxicology
and adjunct faculty in the Department of Environmental and Molecular Toxicology at Oregon State University. Dr. Waters received a PhD in biochemistry from the University of Wisconsin.
Barbara Wetmore is a senior research investigator at The Hamner Institutes for Health Sciences. Her
research focuses on integration of predictive modeling tools with high-throughput screening and other in
vitro strategies to address issues of importance in chemical safety and risk assessment. Other research
110
Prepublication Copy
Appendix A
interests have been the application of genomic and proteomic tools to inform chemical mode-of-action
assessments and biomarker discovery. She is vice-president-elect of the Society of Toxicologys In Vitro
and Alternative Methods Specialty Section, and she has served as a study-section reviewer for the US Environmental Protection Agency and as an expert for the European Union Reference Laboratory for Alternatives to Animal Testing (EURL-ECVAM). Dr. Wetmore received her PhD in toxicology from North
Carolina State University.
Yvonne Will is a senior director and the global science and technology lead for drug safety at Pfizer. Her
research interests include mitochondrial, mechanistic, and cell-based toxicity assessment; drug-induced
organ toxicities; and alternative in vitro and in vivo models. Dr. Will is the president of the Society of
Toxicologys Drug Discovery Specialty Section and a member of the steering committee of the Technology Industry Consortium. She also serves on the editorial board of Current Protocols in Toxicology (John
Wiley and Sons) and Applied In Vitro Toxicology (Mary Ann Liebert, Incy). Dr. Will received her PhD in
biochemistry and biophysics from Oregon State University.
Prepublication Copy
111
Appendix B
Available Data or Databases
There are many sources of acute-toxicity data. They include peer-reviewed literature and secondary
sources, such as The Merck Index (ONeil 2013). Several acute-toxicity databases can be easily searched
by using chemical name, CAS Registry number, chemical structure, and other identifiers, and some can
also be searched on the basis of the type of study, toxic effect, species, sex, dose, exposure duration, and
route of exposure. Some databases will tabulate the results of a search. Comprehensive overviews of major acute-toxicity databases are available (Tsakovska et al. 2006; Lapenna et al. 2010). Brief overviews of
a few are described here.
Several US federal agencies maintain databases of toxicity data. The public version of the US Environmental Protection Agency (EPA) Toxicity Reference Database (ToxRefDB) contains data from chronic, subchronic, developmental, and reproductive studies. It is also linked to the agencys ToxCast program. The EPA Aggregated Computational Toxicology Resource (ACToR) includes acute-toxicity data
that are compiled from the Integrated Risk Information System (IRIS), Organisation for Economic Cooperation and Development (OECD) Summary reports, and Agency for Toxic Substances and Disease
Registry documents. Unlike ToxRefDB, the ACToR database is not directly linked to acute-toxicity information on ToxCast or Tox21 chemicals. EPA also maintains the Toxic Substances Control Act Inventory and the SUPERLIST set of regulatory resources.
The National Library of Medicine (NLM) manages a network of databases called TOXNET,
which makes it possible to search for acute-toxicity information that is available in the Hazardous Substances Data Bank (HSDB). The most convenient means of accessing acute-toxicity information is
through a chemical search that uses ChemIDplus, which has direct links to resources in NLM, other federal agencies, states, and scientific sites. The NLM databases contain records for more than 400,000
chemicals. NLM also maintains Web-site links at http://sis.nlm.nih.gov/chem/alllocators.html to other
databases, such as the Canadian Domestic Substances List, the European Inventory of Existing Commercial Substances, the FDA Drugs@FDA system, and databases of the International Agency for Research
on Cancer.
Several commercially available databases are also available. For example, Leadscope, Inc. markets a
toxicity database that contains nearly 180,000 chemical structures and over 400,000 toxicity-study results
derived from the US Food and Drug Administration Priority-based Assessment of Food Additives
(PAFA) Database, the National Toxicology Program Chronic Database, the Registry of Toxic Effects of
Chemical Substances (RTECS), and the DSSTox Carcinogenicity Potency Database (CPDB) (Leadscope
2012). Acute-toxicity data related to multiple exposure routes are available in the PAFA database and
RTECS.
Several international databases are available. A multinational OECD database, eChemPortal1, is a
no-cost publicly available acute-toxicity database that can be searched by using a variety of chemical
identifiers. One of its strengths is that it includes classification results based on the Globally Harmonized
System of Classification and Labelling of Chemicals (GHS). The European Chemical Agency also man1
See http://www.echemportal.org/echemportal/substancesearch/substancesearchlink.action.
112
Prepublication Copy
Appendix B
ages an electronic database derived from Registration, Evaluation, Authorisation, and Restriction of
Chemicals (REACH) registration dossiers for chemical substances manufactured or imported in Europe.
The OECD QSAR Toolbox is a software tool that facilitates the development, evaluation, justification,
and documentation of chemical categories for read-across. It contains regulatory inventories, toxicity data, and chemical-structure information that encode structureactivity relationship information.
REFERENCES
Lapenna, S., M. Fuart-Gatnik, and A. Worth. 2010. Review of QSAR Models and Software Tools for Predicting Acute
and Chronic Systemic Toxicity. EUR 24639. Luxembourg: Office of the European Union [online]. Available:
http://publications.jrc.ec.europa.eu/repository/bitstream/JRC61930/eur_24639_en.pdf [accessed March 30, 2015].
Leadscope. 2012, Comprehensive Source for Toxicity Data [online]. Available: http://www.leadscope.com/tox
icity_database/ [accessed March 30, 2015].
ONeil, M.J. 2013. The Merck Index: An Encyclopedia of Chemicals, Drugs, and Biologicals, 15th Ed. Cambridge:
Royal Society of Chemistry.
Tsakovska, I., I. Lessigiarska, T. Netzeva, and A.P. Worth. 2006. Review of (Q)SARs for Mammalian Toxicity.
EUR 22846 EN. Luxembourg: Office of the European Union.
Prepublication Copy
113

(Modern Toxo Approaches

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(Modern Toxo Approaches

Uploaded by

Copyright:

Available Formats

This PDF is available from The National Academies Press at http://www.nap.edu/catalog.php?

Application of Modern Toxicology Approaches for Predicting

Committee on Predictive-Toxicology Approaches for Military Assessments

Visit the National Academies Press online and register for...

10% off print titles

Copyright National Academy of Sciences. All rights reserved.

APPLICATION OF MODERN TOXICOLOGY

Committee on Predictive-Toxicology Approaches for

Copyright National Academy of Sciences. All rights reserved.

THE NATIONAL ACADEMIES PRESS

500 Fifth Street, NW

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

COMMITTEE ON PREDICTIVE-TOXICOLOGY APPROACHES FOR

Copyright National Academy of Sciences. All rights reserved.

BOARD ON ENVIRONMENTAL STUDIES AND TOXICOLOGY1

Copyright National Academy of Sciences. All rights reserved.

OTHER REPORTS OF THE

Copyright National Academy of Sciences. All rights reserved.

State and Federal Standards for Mobile-Source Emissions (2006)

Copyright National Academy of Sciences. All rights reserved.

BOARD ON LIFE SCIENCES

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

CONCEPTUAL FRAMEWORK AND PRIORITIZATION STRATEGY ......................................... 13

NONTESTING APPROACHES RELEVANT TO PREDICTION OF ACUTE

ASSAYS FOR PREDICTING ACUTE TOXICITY .............................................................................. 50

INTEGRATION AND DECISION-MAKING FOR PREDICTIVE TOXICOLOGY ........................ 77

Copyright National Academy of Sciences. All rights reserved.

LESSONS LEARNED AND NEXT STEPS ............................................................................................ 93

A BIOGRAPHICAL INFORMATION ON THE COMMITTEE ON

AVAILABLE DATA OR DATABASES ................................................................................................ 112

Copyright National Academy of Sciences. All rights reserved.

Illustration of a general approach to integration and decision-making for applying predictive

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

APPLICATION OF MODERN TOXICOLOGY

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

The verbatim statement of task is provided in Chapter 1 of this report.

BOX 1-1 Statement of Task

BOX 2-1 Definitions of Relevant Terms

Conceptual Framework and Prioritization Strategy

Examples of in vitro Assay Approachesc

Chemical or Biological Agent

Altered axonal transport

Tubulin polymerization assessed with flow

Altered impulse conduction by axonal

Blocking of Na+ ion channel

Cell-based assays of the membrane potential that

Reduced precursor availability or

Inhibition of acetylcholine uptake Vesamicol

PC12 cell-based microelectrode assay (Cui et al.

Altered neurotransmitter release

PC12 cell-based system for in vitro measurements

Opioids, benzodiazepines, nicotine,

Review of selected methods to assess receptor

Nerve gas agents

Altered electrical conduction of heart or

Assessment of altered cardiomyocyte contraction